Ecological Research Series
TOXICITY TO FISH OF
CYANIDES AND RELATED COMPOUNDS
A Review
Environmental Research Laboratory
Office of Research and Development
U.S. Environmental Protection Agency
Duluth, Minnesota 55804
-------
RESEARCH REPORTING SERIES
Research reports of the Office of Research and Development, U.S. Environmental
Protection Agency, have been grouped into five series. These five broad
categories were established to facilitate further development and application of
environmental technology. Elimination of traditional grouping was consciously
planned to foster technology transfer and a maximum interface in related fields.
The five series are:
1. Environmental Health Effects Research
2. Environmental Protection Technology
3. Ecological Research
4. Environmental Monitoring
5. Socioeconomic Environmental Studies
This report has been assigned to the ECOLOGICAL RESEARCH series. This series
describes research on the effects of pollution on humans, plant and animal
species, and materials. Problems are assessed for their long- and short-term
influences. Investigations include formation, transport, and pathway studies to
determine the fate of pollutants and their effects. This work provides the technical
basis for setting standards to minimize undesirable changes in living organisms
in the aquatic, terrestrial, and atmospheric environments.
This document is available to the public through the National Technical Informa-
tion Service, Springfield, Virginia 22161.
-------
EPA-600/3-76-038
April 1976
TOXICITY TO FISH OF
CYANIDES AND RELATED COMPOUNDS
A Review
by
Peter Doudoroff
Department of Fisheries and Wildlife
Oregon State University
Corvallis, Oregon 97331
Grant No. R-802459
Project Officer
Donald I. Mount
Environmental Research Laboratory
Duluth, Minnesota 55804
U.S. ENVIRONMENTAL PROTECTION AGENCY
OFFICE OF RESEARCH AND DEVELOPMENT
ENVIRONMENTAL RESEARCH LABORATORY
DULUTH, MINNESOTA 55804
-------
DISCLAIMER
This report has been reviewed by the Environmental Research Laboratory,
U.S. Environmental Protection Agency, and approved for publication.
Approval does not signify that the contents necessarily reflect the
views and policies of the U.S. Environmental Protection Agency, nor
does mention of trade names or commercial products constitute endorse-
ment or recommendation for use.
11
-------
ABSTRACT
The world literature on the toxicity to fish of simple and complex
cyanides, nitriles, cyanogen chloride, thiocyanates, and cyanates is
reviewed critically and interpretively. Differently determined limits
of toxicant concentrations tolerated by various fishes are compared, and
their variation with exposure time, the pH, temperature, and dissolved
oxygen and mineral content of the water, body size, age, acclimation,
etc., is examined. Interactions of free cyanide with other toxic water
pollutants also are considered. Available data on effects of sublethal
levels of free cyanide on growth, food consumption and utilization,
swimming ability, behavior, etc., and observations on avoidance reac-
tions of fish to the toxicant are summarized and their ecological sig-
nificance is discussed. After a brief introduction to the chemistry
of complex metallocyanides and their behavior in dilute solutions, the
acute toxicity of the solutions is thoroughly considered and related to
concentrations of their identifiable components. The dominant role of
molecular hydrocyanic acid produced by dissociation or photolysis of the
metallocyanide complexes as a lethal agent responsible for the toxicity
of most of the toxic solutions tested is given particular attention; the
relative toxicity of complex metallocyanide ions also is considered.
Some conclusions regarding acceptable concentrations of free cyanide in
receiving waters are presented.
This report was submitted in fulfillment of Grant Number R-802459 by
Oregon State University under the (partial) sponsorship of the Environ-
mental Protection Agency. Work was completed as of September 1975.
iii
-------
CONTENTS
Sections Page
I Conclusions 1
II Introduction 4
General 4
Terminology and Analytical Methods 6
Nomenclature 10
III Lethal Toxicity of Free Cyanide 12
Influence of pH, and the Relative Toxicity of HCN
and the CN Ion 12
Results of Constant-Flow Tests (General) 14
Results of "Static" Tests (General) 20
Relations Between Cyanide Concentration and
Survival Time 30
Influence of Temperature 32
Influence of Dissolved Oxygen 42
Influence of Water Salinity and Hardness 45
Effects of Acclimation to Cyanide 48
Resistance in Relation to Body Size and
Physiological State 50
Interactions of Free Cyanide with Other Poisons 54
Antagonistic Action of Thiosulfate 60
Field Observations of Cyanide-Caused Fish
Mortalities 61
-------
Sections Page
IV Sublethal Toxicity of Free Cyanide and Avoidance
Reactions of Fish 64
Effects on Swimming Ability 64
Effects on Growth, Food Consumption, and Food
Utilization 67
Effects on Embryonic Development, Respiration,
and Heart Beat 70
Tests for Other Nonlethal Injury 74
Avoidance Reactions 78
V Toxicity of Complex Cyanides 83
General, Chemical Background 83
Toxicity of the Metallocyanide Complexes in
General 90
Zinc-Cyanide and Cadmium-Cyanide Complexes 96
Nickel-Cyanide Complex 101
Silver-Cyanide Complex 113
Copper-Cyanide Complexes 120
Iron-Cyanide Complexes 126
VI Toxicity of Other, Related Compounds 135
Nitriles 135
Cyanogen Chloride 140
Thiocyanates 141
Cyanates 143
VII References 144
VI
-------
SECTION I
CONCLUSIONS
The observed toxicity to fish of most of the tested, acutely toxic
solutions of both simple and complex cyanides, and also those of
acetaldehyde cyanohydrin (lactonitrile), is attributable very pre-
dominantly, or almost entirely, to the presence of molecular
(undissociated) hydrocyanic acid, HCN, whose concentrations are
reliably measurable. The toxicity of the cyanide ion, CN , which is
a minor component of the so-called free cyanide (HCN + CN~) in
polluted waters that are not exceptionally alkaline, is indeterminable
but of little importance. Also undetermined but clearly of no impor-
tance is the toxicity of zinc-cyanide and cadmium-cyanide complex
anions, which are almost totally dissociated in very dilute but highly
toxic solutions of the complexes. The acute toxicity of more stable
silver-cyanide and cuprocyanide complex anions is much less than that
of molecular HCN but is not negligible; these ions can be the principal
toxicants even in some very dilute solutions. The much lower toxicity
of the ferrocyanide and ferricyanide complex ions, which are complexes
of high stability but subject to extensive and rapid photolysis,
yielding free cyanide, on exposure to direct sunlight, and also of the
nickelocyanide complex ion, is not likely ever to be of any practical
importance.
\
The chronic toxicity of the metallocyanide complexes has not yet been
investigated. One can reasonably conclude, nevertheless, that in the
absence of important toxic pollutants other than simple and complex
-------
cyanides, the determined molecular HCN content of a cyanide-polluted
water is a fairly reliable index or measure of its toxicity to fish,
except under some rather unusual circumstances that generally can be
readily recognized. The relatively high degree of dissociation of
moderately stable cyanide complexes, such as the nickelocyanide
complex, in exceedingly dilute solutions in which molecular HCN levels
are low but are likely to have serious sublethal effects on sensitive
fishes should be noted in this connection. Also to be borne in mind
is the high toxicity of some of the metal ions. Total (free an^
complexed) cyanide concentrations are not toxicologically meaningful.
In view of present availability of reliable analytical methods for the
determination of low-level molecular HCN and free cyanide in polluted
waters, there appears to be little justification for continued
reliance only on total cyanide determinations in the evaluation and
control of water pollution with cyanides.
Free cyanide or molecular HCN concentrations as low as 0.01 mg/1 can
rapidly and lastingly impair the swimming ability of salmonid fishes
in we11-oxygenated water. Lethal threshold concentrations may be as
low as 0.02-0.025 mg/1 at very low temperatures, though apparently
they are generally above 0.05 mg/1 under favorable conditions. The
susceptibility of these and other fishes to cyanide poisoning is
markedly increased at low oxygen concentrations. Clearly, free cyanide
concentrations above 0.005 mg/1 cannot be always entirely harmless to
the salmonids and other very sensitive fishes, and only much lower
levels may be truly safe concentrations in most waters in which such
susceptible forms must be fully protected from any possible injury by
toxicants.
On the other hand, there is, as yet, no evidence that even persistent
free cyanide or HCN concentrations not far exceeding 0.025 mg/1, in
waters not otherwise seriously polluted, are incompatible with the
-------
persistence of valuable fisheries. No conspicuous impairment by such
concentrations of the swimming ability, growth, embryonic and larval
development, behavior, etc., of relatively resistant species, for
which the lowest experimentally determined lethal threshold concentra-
tions in well-oxygenated water are well above 0.10 mg/1, has been
demonstrated. Such resistant, warm-water species can be the species
of major commercial or recreational importance inhabiting some waters
subject to cyanide pollution, waters whose fisheries may not, for
social and economic reasons, merit protection of the highest degree.
Free cyanide concentrations as high as 0.05 mg/1 seem to be too close
to reported lethal threshold concentrations even for moderately
resistant fishes and to sublethal levels clearly harmful to such
species to be judged acceptable in any waters whose fisheries are to
be afforded more than minimal protection.
The ecological significance of observed metabolic disturbances in
fishes exposed to cyanide solutions and of the impairment of their
swimming ability, as well as of avoidance reactions to the poison
seen in laboratory tests, is still unclear. Much more information
regarding the effects on fish of sublethal concentrations of cyanides
than is now available, especially effects under nearly natural condi-
tions, obviously is needed. Information concerning sublethal
injurious effects of other, related compounds, such as cyanogen
chloride, whose acute toxicity is about as great as that of free
cyanide, and the typical nitriles, whose acute toxicity is less, but
which, unlike free cyanide, evidently can be important accumulative
poisons, is totally lacking.
-------
SECTION II
INTRODUCTION
GENERAL
This is a comprehensive, critical, and interpretive review of the
voluminous world literature on the toxicity to fish of the simple and
complex cyanides and of related compounds, such as the nitriles,
thiocyanates, and cyanogen chloride. Avoidance reactions of fish to
the cyanides also are considered, together with other sublethal effects
of these toxicants. A need for such a review has become increasingly
apparent with increase in number of pertinent publications, in which
some strikingly divergent or seemingly contradictory findings and
evidently erroneous and misleading conclusions have been reported.
Recently intensified efforts strictly to regulate, largely for the
protection of fisheries, the disposal of industrial wastes containing
cyanides have been hampered not only by a lack of much of the needed
information, but also by frequent misunderstanding or ignorance of
pertinent information already available. The present review should be
helpful in this connection and also in the planning of additional
research, which is certainly needed and may be stimulated by definition
of the major, still unsolved or incompletely solved problems. Litera-
ture antedating 1915 is not considered.
Cyanides, whose careless discharge into public waters by electroplaters
and others has in the past caused many disastrous fish mortalities, are
still among the important water pollutants endangering fish and other
-------
aquatic life. They occur in waste waters from metal-finishing and
metallurgical (e.g., gold and silver ore reduction) plants, steel mills,
petroleum refineries, gas plants, and many other industrial sources.
Improved waste disposal controls have reduced the frequency of sporadic,
gross pollution of waters with cyanides (strong wastes, etc.) attrib-
utable simply to carelessness or ignorance, but less obvious, chronic
pollution resulting from continuous discharges of waste waters with
relatively low concentrations of the toxicants is still common.
Definition of maximum concentrations of cyanides that are harmless to
fish life or acceptable in receiving waters whose fish populations are
to be protected so that fish production will not be seriously impaired
is essential to sound waste disposal and water quality management
wherever these toxic compounds can be important pollutants.
Unfortunately, most of the pertinent literature has to do with lethal
effects and tolerance limits, and the available information on sublethal
effects of cyanides is still scanty. Knowledge of the tolerance limits
alone obviously is not a sufficient basis for any definite conclusions
as to limits of entirely harmless concentrations of toxicants. A
brief summary of the lethality data would surely suffice if this
information could be useful only in explaining or preventing fish
mortalities. However, comparison of such data obtained in tests with
different species of fish or under varying experimental conditions
reveals differences in susceptibility of the test animals to lethal
effects of the toxicants that can be reasonably expected to be usually
accompanied by similar differences in susceptibility to sublethal
injury. And much of what has been learned recently about the relative
acute toxicity of the different components of solutions of complex
cyanides (i.e., about the causative agents of the rapidly lethal
toxicity of these solutions) is surely pertinent to sublethal or
chronic toxicity problems also. Results of experiments on the toxicity
of the complex cyanides have been often misinterpreted in published
-------
literature, largely because of insufficient understanding of the
complicated chemistry of dilute solutions of these compounds. Mis-
understanding of the significance of these data has been reflected in
illogical and confusing waste disposal regulations or effluent and
water quality standards that have been proposed or adopted by regula-
tory agencies and can be either excessively or insufficiently restric-
tive or demanding under different circumstances. Widespread adoption
of wastewater treatment that removes nearly all of the free cyanide
renders increasingly important understanding of the toxicological
properties and of the chemical behavior in receiving waters of any
remaining cyanide complexes. For the above reasons, much attention
is given in this review to some comparative lethality studies and
especially to the available data on the acute toxicity of the complex
cyanides and to their interpretation in chemical terms. It is my
hope that future research in the sublethal toxicity of the cyanides
and related compounds thus will be greatly facilitated.
TERMINOLOGY AND ANALYTICAL METHODS
In aqueous solutions of cyanides, the cyanide group, CN, can exist in
different forms that must be distinguished. To avoid possible mis-
understanding, the pertinent terminology used in this review must be
carefully explained, with some attention to analytical methods that
have been used for determination of cyanide in its different forms.
The designations cyanide ion and CN" often have been used in published
literature synonymously with free cyanide or with total cyanide, both
of which terms will be defined presently. Such inappropriate and
confusing terminology probably is responsible for the error to be
found in the generally authoritative publication "Standard Methods for
the Examination of Water and Wastewater" (American Public Health Asso-
ciation, et al., 1971), which attributes the toxicity of cyanide
-------
solutions to the formation in them of the cyanide ion, CN~. Actually,
for reasons to be fully explained later, the toxicity of cyanide-
polluted waters to aquatic life is referable, as a general rule, mostly
or entirely to molecular hydrocyanic acid, HCN, and not to the CN~ ion,
although the latter form of cyanide probably is not without toxicity.
In this review, the term cyanide ion is used in referring only to the
simple, free anion CN~ per se, excluding all other cyanide species.
In my opinion, this is the only strictly correct and toxicologically
sound use of the term.
Hydrocyanic acid, HCN, ionizes in some degree in aqueous solutions,
yielding cyanide ion, the extent of the ionization depending on the pH.
That HCN which is not ionized but is present in the form of uncharged,
intact molecules only, is usually best referred to explicitly as
molecular HCN, hydrocyanic acid, or hydrogen cyanide to avoid mis-
understanding (see the next paragraph), but here it is sometimes
referred to simply as HCN for brevity where the intended meaning is
clear enough.
The term free cyanide is used in referring to both the CN~ ion and the
molecular HCN present in a solution, considered together and without
regard to their sometimes multiple sources, which may be simple, alkali
cyanides or metallocyanide complexes (complex anions) that dissociate
or decompose in varying degrees. Only the weight of the cyanide group
or radical, CN, usually is considered in reporting the free cyanide
concentration in an experimental solution, but sometimes the equivalent
weight of HCN has been reported instead, HCN being almost always the
predominant component. The latter practice cannot be seriously mis-
leading when the mole ratio of CN~ to HCN is very small, as it usually
is, and it should be noted that the difference in molecular weight
between HCN and CN is too small to be of any real consequence in this
connection. The expression "free cyanide as HCN" signifies that CN" is
included.
7
-------
Complexed cyanide is only that cyanide which, is actually bound up in a
complex or complexes and does not include that which has been liberated
through dissociation or decomposition of a complex, forming HCN or CN
ion.
The term total cyanide has been used in the literature in somewhat
different senses. When used by chemists concerned with water quality,
it usually embraces all cyanide species or groups determinable by a
particular, selected analytical method involving acid distillation, in
the course of which all or most of the complexed cyanide usually is
liberated as HCN and trapped for measurement as CN ion. Although the
term implies no exclusion, actually not all of the cyanide groups
present in a sample may be included in the measurement, and values
obtained for the same sample by different, widely used and approved
procedures may differ considerably. Some cyanide not initially present
may even be generated during the distillation. In this review, the
term total cyanide is used as a designation for both free cyanide and
complexed cyanide considered together and without regard to amounts
recoverable or measurable by any analytical method. Thus, it is used
in reporting amounts of cyanide as CN known to have been added to
water in one form or another in preparing experimental solutions in
which some of the cyanide was known to have been complexed. It may be
qualified by words such as "initial", "added", or "introduced" when
there are reasons for believing that a considerable portion of the
added cyanide was or may have been lost to the atmosphere, destroyed by
chemical reaction, or otherwise eliminated during or before a test of a
solution for toxicity.
Standard analytical methods for the colorimetric determination of total
cyanide in water are well known and have long been used extensively in
connection with water pollution control, but they have only very
limited applicability to toxicological research. On the other hand,
8
-------
only recently (i.e., in the past 15 years) has there been much interest
and progress in the development of reliable and sufficiently sensitive
analytical methods for the determination of free cyanide, molecular
HCN, or CN ion in waters containing cyanide complexes of widely
varying stability. Although several such methods have now been
developed and published, they are still little known and apparently
have been profitably employed in the water quality field only by a few
investigators engaged in research in fish toxicology. There appears to
have been a curious indifference to these methodological advances or
resistance to their practical application, even though it has long been
obvious that the total cyanide determinations made and reported
routinely in the past are often insufficiently instructive, if not
meaningless, as measures of water quality.
Most of the recently developed methods for determination of molecular
HCN are essentially modifications of a crude colorimetric method
published long ago by Worley and Browne (1917). Each involves distri-
bution of the HCN between water (the sample) and dispersed air or
nitrogen, trapping of the displaced HCN in one way or another, and its
subsequent measurement as HCN or as CN ion in an alkaline solution.
The final measurement is done by gas-liquid chromatography, using a
thermal conductivity detector (Schneider and Freund, 1962) or, for
increased sensitivity, a flame ionization detector (Claeys and Freund,
1968), polarographically (Nelson and Lysyj, 1971), or colorimetrically
(Broderius, 1973). Accuracy of such methods depends on the use of
water samples large enough and volumes of air or nitrogen small enough
to ensure that only a very small fraction of the HCN initially present
in each sample will be displaced and there will be no material change
in pH of the sample due to displacement of carbon dioxide. Material
disturbance of existing equilibria thus can be avoided. Somewhat
different in principle is a method that involves colorimetric measure-
ment of the transfer of HCN from a water sample to a sodium hydroxide
-------
solution by diffusion through air in a sealed container (Great Britain,
Ministry of Technology, 1967; Brown, Shurben, and Shaw, 1970).
Decidedly different is the methylchloroform extraction method described
by Montgomery, Gardiner, and Gregory (1969), which is convenient to use
in the field and appears to be sound in principle but has not been
shown to be sufficiently reliable. For reasons not known, much dis-
agreement between HCN concentrations found by the authors in solutions
containing cyanide complexed with nickel and the corresponding,
computed equilibrium levels of molecular HCN in these solutions has
been noted (Broderius, 1973).
Cyanide ion and free cyanide concentrations in waters of known pH can
be readily derived from determined levels of molecular HCN. Even the
most refined electrochemical methods for determination of cyanide ion
that have been recently developed probably do not have the sensi-
tivity requisite for detection, in waters that have not been rendered
highly alkaline, of the very low levels of free cyanide that often
need to be measured.
NOMENCLATURE
Common names are used mostly in this review for species of fish that
investigators have used as test animals, but scientific names also are
given. Currently preferred or accepted common and scientific names
are used in place of some of those given by authors whose findings are
reported. A scientific name, once given, is not repeated when the same
species, identified by common name, is mentioned more than once in the
same paragraph or in successive paragraphs. Moreover, for a small
number of species that have been much used as test animals, only the
common names usually are given, for frequent repetition of the
scientific names of these well-known species after their first mention
is deemed unnecessary and would be wasteful of space. A list of the
10
-------
common and scientific names of these repeatedly mentioned species
follows:
Bluegill Lepomis macrochirus
Brovm trout Salmo trutta
Fathead minnow Pimephales promelas
Guppy Poecilia reticulata (Lebistes reticulatus)
Rainbow trout Salmo gairdneri
Threespine stickleback Gasterosteus aculeatus
11
-------
SECTION III
LETHAL TOXICITY OF FREE CYANIDE
INFLUENCE OF pH, AND THE RELATIVE TOXICITY OF HCN AND THE CN~ ION
The distribution of free cyanide between its two different forms, that
is, the mole ratio of the CN~ ion to molecular HCN, depends on the so-
called hydrogen ion concentration [activity of hydronium ion, HgO"1",
commonly still designated for convenience by the symbol H ). Pertinent
equilibrium equations are:
[CN~] [CN~] K
= Ka , or
a
[HCN] [HCN] [H+]
where Ka is the ionization constant of HCN, which increases markedly
with increase of temperature and is about 4-6 x 10~1D at 20-25° C, and
the bracketed symbols represent molar concentrations (or, strictly
speaking, activities) of the indicated species or solution components.
The exact value of [CN~] in any cyanide solution thus can be derived
from a known value of [HCN], and vice versa, and both can be derived
from their sum, a known concentration of free cyanide, if the pH of the
solution, which is the negative logarithm of [H+], and the appropriate
value of Ka are exactly known. At pH 8.3, which is not often exceeded
in natural, surface waters, only about 9 percent (not more than 11%)
of any free cyanide occurs as cyanide ion when the temperature of the
water is between 20 and 25° C, and the percentage declines steeply
12
-------
with reduction of the pH, becoming only about 0.5 percent at neutrality
(pH 7.0). Even at the quite unusual pH of 9.0, some two-thirds of the
free cyanide is in the form of molecular HCN. Thus, even if the two
forms of free cyanide were of equal toxicity, most or nearly all of the
toxicity of waters polluted mainly with free cyanide would be usually
due to their molecular HCN content and not correctly attributable to
the presence of cyanide ion.
The acute toxicity to fish of solutions of simple alkali cyanides, or
of free cyanide, has been found to decrease with increase of the pH of
the solutions, and it has been concluded, therefore, that molecular
HCN is more toxic than the CN~ ion. In a small series of tests per-
formed by Wuhrmann and Woker (1948), the average time to lasting loss
of equilibrium of the chub (cyprinid) Squalius cephalus in sodium
cyanide (NaCN) solutions of the same cyanide content (0.66 mg/1 as CN)
and varying pH increased from about 52 minutes to 70 and 94 minutes with
increase of pH from about 7.6-7.7 to 8.12 and 8.84, respectively. How-
ever, it declined again somewhat, to 73 minutes, with further increase
of the pH to 9.5. The decrease of toxicity with increase of pH was
attributed by the authors to the decrease of the concentration of
molecular HCN, and the increased toxicity at the highest pH to the
additional harmful action of the very high pH itself. Bridges (1958)
reported that fish (largemouth bass, Micropterus salmoides, green sun-
fishj Lepomis cyanellus, and others) were more resistant to NaCN (1.0
mg/1 as NaCN, or 0.53 mg/1 as CN) at pH 8.4 and 8.9 in aquaria and at
pH 9.7 in a pond treated with the chemical than they were at much lower
levels of pH. In aquarium tests with only a few fish, largemouth bass
lived for about 80 minutes at pH 8.9 but were killed in about 50 min-
utes, on the average, at pH 7.0-7.3. Fish began to come to the surface
only about 30 minutes after application of the poison to the pond with
pH 9.7 and within 15 minutes in a comparable pond with about the same
surface temperature. Water temperatures were around 25° C in the
aquarium tests and 17° C in the comparable ponds.
13
-------
The above, limited observations cannot be said to be entirely conclu-
sive. However, they are in agreement with some earlier observations on
the toxicity of HCN to aquatic organisms other than fish, and also with
observations on the toxicity to fish of salts of another weak acid,
hydrogen sulfide, and the weak base ammonia in solutions of varying pH
(Doudoroff and Katz, 1950). In each case, the undissociated or molecu-
lar form of the acid or base was found to be much more toxic than the
acid anion or base cation. The greater toxicity of the molecular form
is attributable to the relative ease with which small, uncharged
molecules penetrate into the blood and other tissues of organisms,
whose external membranes are relatively impermeable or less permeable
to the charged ions. Internally, the chemically reactive cyanide ion
must be the active agent of cyanide poisoning, just as the hydronium
ion, which also does not penetrate living membranes readily, must be
the active agent of poisoning of fish with carbon dioxide and carbonic
acid (HoCO?), whose anion is harmless. Alexander, Southgate, and
Bassindale (1935) reported that no effect of pH variations within the
range of 6.0 to 8.5 on the toxicity of KCN solutions (0.3 mg/1 as CM)
to rainbow trout was seen. However, the failure to observe differences
of overturning time in tests with limited numbers of fish may well have
been due to insufficient change of the rapidly lethal molecular HCN
content of test solutions with a change of pH within the experimental
range at the low test temperature (7-8° C).
RESULTS OF CONSTANT-FLOW TESTS (GENERAL)
The acute or more or less rapidly lethal toxicity of simple alkali
cyanides to fish at ordinary experimental temperatures and under other-
wise favorable conditions has been studied by a large number of inves-
tigators. Free cyanide is not a persistent toxicant, and so is lost
fairly rapidly from unrenewed, standing test solutions in which fish
are held. Therefore, the results of only those fairly prolonged
14
-------
toxicity tests during which, the test solutions are continuously renewed
(i.e., so-called constant-flow or continuous-flow bioassays) can be
entirely reliable. Such tests, with complete immobilization or perma-
nent overturning (loss of equilibrium) of the fish taken as the end-
point of "survival", have been performed only with a rather small number
of fish species of limited variety. The duration of a number of these
tests was as long as 3 days or much longer, but others lasted for only
about 1 day or less. Average or median periods of survival of the fish
at tested concentrations of the toxicants, or percentages of test.
animals surviving at these concentrations for various, sometimes quite
prolonged, exposure periods have been determined. From recorded sur-
vival percentages, median tolerance limits (median lethal concentra-
tions, or toxicant concentrations at which just 50 percent of the test
animals are able to survive) for the periods of actual exposure of the
fish to test solutions have been derived by interpolation. Also, lethal
threshold concentrations (incipient lethal levels, or concentrations
nearly or barely tolerable for individuals of average resistance when
exposure thereto is indefinitely prolonged) often have been estimated
by extrapolation, i.e., derived from observed relationships between
concentration and average or median survival time, using graphical or
computational methods believed to be appropriate. Some of the estimates
(indicated threshold levels) have been based on results of tests of
rather short duration only. Most of the estimated lethal threshold con-
centrations and reported or indicated median tolerance limits for
exposure periods of about 3 days or longer have been found to fall with-
in the range of 0.05 to 0.16 mg/1 as CN when results of constant flow
tests only were considered, but a few of the values are well below or
above this range. Potassium cyanide (KCN) was used in most of the tests.
Karsten (1934) reported that all of eight brook trout, Salveljnus
fontinalis, about 15 cm long died within 136 hours at a free cyanide
concentration (in KCN solutions) of 0.05 mg/1 as CN, and all of six
15
-------
survived for 27 days at 0.02 mg/1 as CN. However, Neil (1957) found a
KCN solution with a concentration of 0.05 mg/1 as CN (amount of cyanide
added to water) to be nonlethal to brook trout 10-13 cm long in an
exposure period of 40 days at least; only five of ten brook trout had
turned over permanently after 87 hours at a concentration of 0.08 mg/1
as CN. These results agree well with those obtained with young rainbow
trout, Salrno gairdneri, by Herbert and Merkens (1952), who reported an
average survival time of 74 hours at a KCN concentration of 0.07 mg/1
as CN. They agree also with results obtained in experiments with young
brown trout, Salmo trutta, by Burdick, Dean, and Harris (1958), who
estimated the lethal threshold concentration for these fish at high
oxygen concentrations to have been about 0.09 mg/1 as CN, and who
observed only a 40 percent mortality after 10 days in one test at a KCN
concentration of 0.07+ mg/1 as CN. Test temperatures were near 9.5° C
in Neil's experiments, 17.5° in those of Herbert and Merkens, whose
rainbow trout averaged a little less than 10 cm in length, and about
15.5° C in the experiments of Burdick, Dean, and Harris, who tested
brown trout averaging less than 5 cm and about 10 cm in length with
quite similar results; the water temperature in Karsten's experiments
is unknown.
At a test temperature of 12-13° C, more recent experiments with young
rainbow trout reported by British investigators (Great Britain, Ministry
of Technology, 1968) have yielded results closely agreeing with the
above-mentioned data on the cyanide tolerance of the same and other
kinds of trout. A lethal threshold concentration of free cyanide in
the vicinity of 0.08 mg/1 as CN (between 0.08 and 0.09 mg/1 as HCN)
seemed to be indicated. However, when tests were performed at much lower
temperatures near 3° C, the results indicated a lethal threshold level
between 0.02 and 0.025 mg/1. Additional tests performed later also
demonstrated greatly increased sensitivity of the trout to free cyanide
at very low temperatures. Thus, 24-hour median lethal concentrations
16
-------
for rainbow trout acclimated for 7 days to a temperature of 3° C
averaged about 0.04-0.05 mg/1 (Great Britain, Department of the Environ-
ment, 1972), suggesting that the lethal threshold concentration for
these fish was probably below 0.03 mg/1. These interesting results will
be more fully considered later, in connection with a general discussion
of the influence of temperature on resistance to cyanide. Data on the
adverse effects of low dissolved oxygen concentrations also will be
reviewed in detail later. Burdick, Dean, and Harris (1958) observed
only a moderate effect of reduction of the oxygen concentration to 5.2
mg/1 on the cyanide tolerance of their brown trout at about 15.5° C.
Under these conditions, the lethal threshold concentration was estimated
by them to Trave been about 0.08 mg/1 as CN.
The lethal threshold concentration of cyanide for smallmouth bass,
Micropterus dolomieui, averaging about 6.3 cm in length, in KCN solutions
at a temperature of 21° C, was estimated by Burdick, Dean and Harris
(1958) to have been about 0.104 mg/1 as CN at high oxygen concentra-
tions, but near 0.086 mg/1 at the moderately reduced dissolved oxygen
concentration of 4.2 mg/1. At the high oxygen concentrations, the 23-
hour median tolerance limit was 0.127 mg/1; this cyanide concentration
did not kill all of the test animals in 3 days. Doudoroff, Leduc, and
Schneider (1966) found the 48-hour and 72-hour median tolerance limits
(50 percent lethal concentrations) of NaCN for young bluegills, Lepomis
macrochirus, at 20° C to be nearly equal, about 0.16 mg/1 as HCN (0.154
mg/1 as CN). Of eight bluegills that survived for 48 hours at test
concentrations of 0.155 and 0.180 mg/1 as HCN, from a total of 20 fish
averaging about 5 cm in length that were tested at these two concentra-
tions, only one died during the following day; the remaining seven were
apparently healthy after the 72-hour exposure. Lipschuetz and Cooper
(1955) reported 0.22 mg/1 as CN to have been the 24-hour median tolerance
limit of KCN for blacknose dace, Rhinichthys atratulus meleagris, about
3.8-7.6 cm in length at temperatures around 21° C and high oxygen con-
centrations.
17
-------
Brockway (1963) found the 24-hour and 48-hour median tolerance limits
of NaCN at 25° C for juvenile cichlids, Cichlasoma bimaculatum, 4-5
months old and weighing about 1.0-1.3 g, to be about 0.18 and 0.135
mg/1 as CN, respectively. In a prolonged test with young of the same
species initially about 2 months old and averaging about 0.5 g in
weight, no deaths were observed during a 60-day exposure of the fish
to a concentration of 0.11 mg/1 as CN. This species of aquarium fish
has been much used in studies of sublethal effects of cyanide poisoning
to be considered later. The highest long-term tolerance limit of free
cyanide for fish that has been evaluated by the constant-flow bioassay
technique and reported in published literature apparently is the
theoretical lethal threshold concentration for guppies, Poecilia
reticulata (Lebistes reticulatus), 1 month old at 24-24.5° C estimated
by Chen and Selleck (1969), namely, 0.236 mg/1 as CN in solutions of
KCN. Anderson (1974), however, found the 96-hour median tolerance limit
of KCN for small, adult, male guppies about 0.1 g in weight at 25° C to
be about 0.147 mg/1, doubtless as CN, although reported (erroneously,
I am sure) as a concentration of KCN. And the median survival time of
guppies at a free cyanide concentration of 0.20 mg/1 as CN and 18° C,
probably in a constant-flow test (bioassay method not stated), had been
previously reported to have been about 80 hours (Great Britain, Depart-
ment of Scientific and Industrial Research, 1956). The latter result
was one obtained in a comparative study of the relative resistance to
free cyanide of ten species of fish, of which the guppy proved to be by
far the most resistant one. Abram (1964) estimated a lethal threshold
concentration of but 0.071 mg/1 as CN for the Harlequin fish, Rasbora
heteromorpha (another aquarium fish); the test conditions were not
specified, but it seems reasonable to assume that the test solutions
were continuously renewed.
The data reported by Renn (1955), obtained also by the constant-flow
bioassay technique, are of limited value and not easily summarized and
18
-------
compared with, the data reported above, because of the manner of their
presentation (mostly graphical), as well as the relatively- short dura-
tion of exposure to the test solutions of the several species of warm-
water fish tested. Renn stated that 100 percent survival for 10 hours
of the white crappie, Pomoxis annularis, proved uncertain Cat 25° C) at
KCN concentrations exceeding 0.03 mg/1 as N (0.056 mg/1 as CN), and
that such survival of none of the tested species was observed at con-
centrations exceeding 0.06 mg/1 as N CO.11 mg/1 as CN). He also found
that 50 percent mortality of bluegills and of redbreast sunfish,
Lepomis auritus, often occurred at KCN concentrations as low as 0.11 to
0.14 mg/1 as CN after exposure periods of only about 150 to 350 minutes
Cat 25° C). These concentrations are lower than the long-term median
tolerance limit (of NaCN concentration at 20° C) for the bluegill
estimated by Doudoroff, Leduc, and Schneider (1966), about 0.154 mg/1
as CN.
Reported cyanide concentrations commonly have been amounts of cyanide
added to the experimental water, and not amounts found by chemical
analysis to have been actually present. When chemical analyses of the
test solutions have been made and the results reported, for comparative
purposes, together with the expected (computed) cyanide concentrations
based on amounts of cyanide added, the analytically determined values
generally have been shown to be, for some undetermined reasons, some-
what lower than the expected concentrations. Estimates of tolerance
limits thus may be often a little too high. For example, Herbert and
Merkens (.1952) found the analytically determined cyanide concentration
in the water to which 0.07 mg/1 of cyanide (as CN) had been added and
in which their rainbow trout survived for 74 hours, on the average, to
be only about Q.06 mg/1. Likewise, Doudoroff, Leduc, and Schneider
(1966) found that the molecular HCN concentrations in fresh, standing
NaCN solutions like those used in their long-term, constant-flow,
toxicity tests, determined by the analytical method of Schneider and
19
-------
Freund (1962), were less than the expected (computed) concentrations by
about 8 percent. The correct 48-hour and 72-hour median tolerance
limits of free cyanide for their bluegills thus may have been near 0.14
mg/1 as CN. Cairns and Scheier (1968), who performed "static" bioassays
only (i.e., tests without continuous renewal of test solutions), also
reported initial cyanide concentrations determined by analysis of their
solutions to have been usually, but not always, decidedly lower than the
expected concentrations computed from the amounts of cyanide added to
their synthetic dilution water (.i.e., a water prepared from distilled
water by adding various chemicals). Analytical errors may well have
been responsible, of course, for some of the differences in question.
Burdick, Dean, and Harris (1958) and Renn (1955) stated that their test
solutions were analyzed for cyanide, and their reported test concentra-
tions may well have been based on the results of these analyses. No
comparisons of the analytically determined concentrations with expected
or computed concentrations were reported by these authors. Neil (1957)
and Lipschuetz and Cooper (1955) analyzed some of their test solutions,
but chose to rely on computed cyanide concentrations in reporting their
test results.
RESULTS OF "STATIC" TESTS (GENERAL)
Because cyanide concentrations in unrenewed test solutions decline fairly
rapidly, tolerance limits determined by so-called "static", or standing
water, toxicity bioassays with no renewal of test solutions tend to be
somewhat higher than those determined by constant-flow tests of equal,
fairly prolonged duration. However, disagreement of the results even
of quite prolonged bioassays. of the two kinds generally has not been
very great. Probably the main reason is that cyanide is a fairly rapidly
acting poison; consequently, the relation between free cyanide concen-
tration and survival time is such that maximum constant concentrations
tolerated for long periods, or indefinitely, are not much lower than
20
-------
those just tolerated for periods of only 10 to 24 hours. Acclimation
of surviving fish exposed to critical levels of the poison may be
partly responsible for long or indefinite survival of these fish at
concentrations that had proved rapidly lethal to other individuals, or
at only slightly lower concentrations, even in constant-flow tests.
Review of the many results of determinations of the toxicity of free
cyanide to fish by simple, static bioassay that are to be found in the
literature does reveal some published tolerance limits that are inor-
dinately high and grossly misleading. For example, Dorier (1952) has
reported that a sodium cyanide concentration of 1/200,000 (.i.e., 5 mg/1
as NaCN, or about 2.7 mg/1 as CN) is lethal to brown and rainbow trout,
and a concentration half as great is dangerous for them. And concen-
trations four times as great, 1/50,000 and 1/100,000 (i.e., 10.6 and
5.3 mg/1 as CN) were said to be lethal and dangerous, respectively, for
the roach, Rutilus rutilus. Concentrations as high as 1/500,000 (about
1.1 mg/1 as CN) and 1/200,000 (2.7 mg/1 as CN) then were indicated to
be harmless to the trout and the roach, respectively. These and other
such underestimates of the toxicity of free cyanide must be attributed
to faulty experimental techniques or, in some cases, to insufficiently
prolonged exposure of the fish to test solutions, I suppose.
On the other hand, there are also some published reports of toxicity of
free cyanide to certain fishes that is much greater than that observed
by several other investigators. For example, Ellis (1937) reported
that KCN concentrations of only 0.1 to 0.3 mg/1 (0.04 to 0.12 mg/1 as
CN) in hard water killed goldfish, Carassius auratus, in 3 to 4 days.
Other pertinent data in the available literature indicate that the gold-
fish is among the fishes least susceptible to cyanide poisoning. Powers
(1917) reported that a 0.12 x 10 M solution of KCN, or about 0.31 mg/1
as CN, killed goldfish in about 2 to 5 days at 21.5° C. And Costa
(1965), who found the 50-hour tolerance limit of NaCN for each of four
21
-------
other fish, species tested by him to be not much greater than 0.15 mg/1
as CN, surprisingly reported a corresponding tolerance limit for the
goldfish (at 17-18° C) greater than 1.5 mg/1 as CN (above 6 x 10~ N) .
At least, that is what is shown in his graph; his statement, on page
58, that "young goldfish survive...in a solution of 0.00001 N for more
than 3 days" is not contradictory but is somewhat incongruous, or not
fully in accord with the data presented graphically- His Figure 6 also
is a little confusing, because the value 5 x 10~5 appears on each scale
of abscissas (concentrations) where the value 5 x 10 obviously belongs
instead. Bridges (1958) found that NaCN at a concentration of 1.0 mg/1
(0.53 mg/1 as CN) killed some of his goldfish only after an exposure of
about 2 days and 0.5 mg/1 killed none in 72 hours at 24-28° C. No
explanation can be offered for the low lethal levels reported by Ellis,
except the possibility of error in the preparation of. test solutions or
of serious, undetected depletion of dissolved oxygen in the solutions.
Fish embryos and young larvae have been found to be far more resistant
to free cyanide than are the fully developed animals. For example,
Cairns, Scheier, and Loos (1965) have reported that the 48-hour median
tolerance limit of KCN for "eggs" (embryos) of the zebra danio,
Brachydanio rerio, was estimated to have been about 11.7 mg/1 as CN at
temperatures near 24° C, whereas the reported, corresponding value for
adults of the same species was 0.49 mg/1 as CN. None of the eggs
became opaque after exposure for 48 hours to a KCN concentration of 10
mg/1 as CN. The embryos tested were at a fairly early stage of
development; although all controls developed eye pigmentation within 24
hours after the beginning of the test, the embryos exposed to the
tolerated, high concentrations of cyanide, which evidently inhibited
their development, did not. Brinley (1930) found that the hearts of
6-day old embryos of the mummichog, Fundulus heteroclitus, exposed at
an unspecified temperature to a M/160 solution of KCN in sea water (162
mg/1 as CN) stopped beating only after about 27 hours. Karsten (1934)
reported that a free cyanide concentration of 3.2 mg/1 as CN, which
22
-------
killed brown trout 20-23 cm long within 7 minutes, was tolerated for
several hours with no evident harmful effect by brown trout sac fry
only a day old. He stated that unhatched and partially hatched trout
eggs hatched successfully and the resulting fry remained alive at this
cyanide concentration. However, Philips (1940) found that embryos of
the cunner, Tautogblabrus adspersus, a marine fish which may be exceed-
ingly susceptible to cyanide poisoning, could not tolerate a concentra-
tion of NaCN as high as 0.65 mg/1 as CN; a concentration half as great
was tolerated for a 27-hour observation period.
Most of the reported tolerance limits of free cyanide for juvenile and
adult fishes exposed to the poison for about 1 day or longer (i.e.,
excluding values that pertain to much shorter exposure periods) that
have been determined by static toxicity bioassays apparently without
any renewal of test solutions fall within the range of 0.1 to 0.3 mg/1
as CN. Such tolerance limits falling within this range have been
reported for many common cyprinids, such as the fathead minnow,
Pimephales promelas (Doudoroff, 1956; Henderson, Pickering, and Lemke,
1961), the European minnow, Phoxinus phoxinus (Costa, 1965), and the
Asiatic, carp-like mrigal, Cirrhina mrigala; (Seth, et_ al_., 1967), for
some generally hardy forms such as young eels, Anguilla japonica and
A. anguilla (Oshima, 1931; Costa, 1965), for various centrarchids, such
as the green sunfish, Lepomis cyanellus, (Lewis and Tarrant, 1960) and
the bluegill, which has been used in many experiments whose results will
be summarized and compared presently, for brown trout young and three-
spine sticklebacks, Gasterosteus aculeatus, (Costa, 1965) and for many
other kinds of fish.
Tolerance limits for warm-water fishes well below 0.1 mg/1 as CN, evalu-
ated by static bioassays, have been reported occasionally. Daugherty
and Garrett (1951) found the 24-hour median tolerance limit of HCN
(actually free cyanide as HCN) for the marine pin perch, Lagodon
23
-------
rhomboides, to be only about 0.07 mg/1, even though their test solu-
tions, at uncontrolled temperatures ranging from 13.7 to 20.4° C, were
continuously aerated with compressed air during the tests. Malacca
(1966) reported the minimal lethal concentration of KCN for the cyprinid
fish Leucaspius delineatus to be 0.06 mg/1 as CN, a concentration which,
according to his table of test results and an accompanying graph,
actually killed these fish in about 2.5 hours at 19.5° C and pH 7.5.
However, Gillar (1962), who performed 72-hour static tests, reported
the tolerance limit of KCN for Leucaspius delineatus to be below 0.3
mg/1, a concentration at which 100 percent of the fish died, but well
above 0.14 mg/1 as CN, a concentration at which he observed no evident
reaction of the fish to the presence of the poison. Thus, he found
this fish to be less sensitive than the roach, Rutilus rutilus, for
which, at the same test temperature (12° C), he determined the average
lethal concentration limit to be 0.145 mg/1 under the conditions of his
tests. On the other hand, he found the lethal concentration limit for
the percid fish Acerina cernua (ruffe) to be less than 0.1 mg/1 as CN,
a concentration at which all of these fish died in the test. And Woker
and Wuhrmann (1950) estimated lethal threshold concentrations of HCN to
be about 0.06 mg/1 for the minnow Phoxinus laevis and about 0.08 mg/1
for the perch Perca fluviatilis (but 0.10 mg/1 for the hardy tench,
Tinea vulgaris, and 0.30 mg/1 for the chub Squalius cephalus). These
estimates were derived by extrapolation, according to an equation deemed
appropriate, probably from results of tests of relatively short duration,
and so should not be regarded as proven thresholds. Such values tend to
be lower, of course, than observed limits of tolerance for exposure
periods that are not very long.
Lethal concentration limits of free cyanide well above 0.3 mg/1 as CN
have been recorded not infrequently for resistant species of fish, even
when the duration of the tests on which the values were based was as
long as 1 day or longer. For example, Wallen, Greer, and Lasater (1957)
24
-------
reported that the 24-hour median tolerance limit of KCN for the mosquito-
fish, Gambusia affinis, in a turbid water at 21-23° C was about 1.6 mg/1
(0.64 mg/1 as CN); the 96-hour tolerance limit was found to be the same.
Schaut (1939) reported data indicating a 24-hour median tolerance limit
of NaCN near 0.75 mg/1 (0.40 mg/1 as CN) for unidentified fresh-water
"minnows (killies)." Wells (1916) found that a 1/25,000 N solution of
KCN (1.04 mg/1 as CN) killed the hardy black bullhead, Ictalurus melas,
only after an exposure of almost 28 hours at an unspecified temperature,
although other species of fish were killed rapidly (in about 1 hour or
less) at this concentration. Nehring (1964) reported that yearling
carp, Cyprinus carpio, about 11 cm long and weighing about 50 g, exposed
to KCN solutions at 16° C, were evidently affected at concentrations of
0.4 mg/1 as CN or more, but died within 3 days only at a concentration
of 0.6 mg/1. This report of high resistance of young carp to cyanide
poisoning is not unique, but Silaichuk (1969) observed death of 40
percent within 24 hours at a concentration of only 0.1 mg/1 as CN.
With the exception of two reports already mentioned and one of a test
in which loss of cyanide from the test solution was known to have been
rapid (Kariya £t al., 1967), I have found no published reports of
ability of fully developed fish to survive for reasonably prolonged
exposure periods in solutions with initial free cyanide concentrations
much above 1.0 mg/1 as CN. Even 0.3 mg/1 may not be truly a nonlethal
concentration for any fish species. But there is much evidence that
some resistant warm-water species can live at free cyanide levels four
or five times as great as levels that are lethal to other, more sensi-
tive species (exclusive even of the cold-water salmonids) under essen-
tially the same conditions. The large apparent differences in resistance
among fish of various kinds tested by different investigators can be
reasonably attributed only in part to variation of experimental methods
and conditions, to which some of the noted, striking discrepancies of
test results clearly must be ascribed. Large interspecific differences
have been reported by some investigators who have tested several or
25
-------
numerous species for comparative purposes, presumably under quite
similar conditions. Some of the comparable but very different results
obtained with different species by the same investigators (Costa, 1965;
Woker and Wuhrmann, 1950) already have been mentioned. Costa1s gold-
fish apparently tolerated for 50 hours a free cyanide concentration about
10 times the maximum concentration tolerated by the more susceptible
warm-water fishes used in his experiments. In Malacca's (1966) experi-
ments, in which the cyprinids Leucaspius delineatus apparently were
killed in about 2.5 hours by only 0.06 mg/1 as CN (as already noted),
Phoxinus phoxinus were killed in about 5 hours by 0.20 mg/1, and the
bitterling, Rhodeus sericeus amarus, only in about 8 hours by 1.16 mg/1,
at temperatures ranging from 18.5 to 22° C and pH 7.5-7.9. In other
tests, median resistance periods at a cyanide concentration of 0.20 mg/1
as CN and 18° C are reported to have been less than 3 hours for the
rainbow trout, about 4, 8, and 12 hours for the Harlequin fish, Rasbora
heteromorpha, the minnow Phoxinus phoxinus, and the zebra danio,
Brachydanio rerio, respectively, and about 80 hours for the guppy, whose
median resistance time at a concentration of 1.0 mg/1 was well over 4
hours (Great Britain, Department of Scientific and Industrial Research,
1956). The nature of the tests (i.e., the method employed) is unknown,
but constancy of cyanide concentrations is suggested by the relations
between the tested concentrations and the median resistance periods
presented graphically. In an unpublished report of an early investiga-
tion (Michigan Department of Conservation, Institute for Fisheries
Research, 1933), in which many species were tested, the following
"toxicity thresholds" of NaCN, here converted to equivalent cyanide con-
centrations as CN, were reported, among others: 0.53 mg/1 as CN for
adult mudminnows, Umbra limi, 0.53 mg/1 or less for young carp; 0.39
mg/1 for mottled sculpin, Cottus bairdi, adults; 0.265 mg/1 for yellow
bullhead, Ictalurus natalis_, fingerlings and for half-grown pumpkinseeds,
Lepomis gibbosus; 0.175 mg/1 for rainbow darter, Etheostoma caeruleum,
adults, for half-grown grass pickerel, Esox americanus vermiculatus,
26
-------
and for rock bass, Ambloplites rupestris, fingerlings; 0.155 mg/1 for
largemouth bass, Micropterus sa.lmoides, fingerlings; and 0.11 mg/1 for
rainbow trout fingerlings.
Losses of free cyanide at different rates from unrenewed test solutions
doubtless have been responsible for much, of the variability of static
test results obtained by different investigators, as well as the dis-
agreement of results obtained by different methods. Doudoroff (1956)
observed that an NaCN solution that killed all of his fathead minnows
within 14 hours when it was fresh killed only 60 percent of minnows
introduced after removal of the dead fish and when the solution was 24
hours old. Minnows introduced into the same solution when it was 96
hours old and the fish surviving in it had been removed were not
noticeably affected at all. Bridges (1958) reported a number of similar
experiments with green sunfish, Lepomis cyanellus; in these experiments,
no fish apparently were placed in the aging solutions before the solu-
tions were tested for toxicity. The solutions, with an initial NaCN
content of 1.0 or 1.5 mg/1 (0.53-0.80 mg/1 as CN), killed all the test
fish after aging for 24 hours, but killed none after aging for 72 hours
in some experiments and only 48 hours in others. The depth of the test
solution in Doudoroff's experiment was about 18 cm; that in Bridges'
experiments was not stated. Cairns and Scheier (1963, 1968) determined
initial and residual cyanide concentrations in their unrenewed test
solutions, which probably were continuously oxygenated by mild,
controlled aeration. They stated that the losses of analyzable cyanide
during the 96-hour experimental period varied from 16 to 62 percent of
the initial concentration (Cairns and Scheier, 1963). However, their
tabulated data seem to show 96-hour losses frequently as great as 87.5
percent and always in excess of 40 percent. The determinations of the
very low residual levels admittedly were not very reliable or precise.
Kariya et al. (1967) observed a very rapid decline of cyanide concentra-
tion in their evidently overloaded test aquaria.
27
-------
Burdick, Dean, and Harris (1958) compared results of parallel static
and constant-flow toxicity bioassays of KCN, using brown trout as test
animals. Although only rather high, rapidly lethal concentrations
were tested, which killed all the fish in less than 4 hours, there was
a pronounced and statistically quite significant difference of the
results obtained by the two methods at the lowest concentration. In
the constant-flow test, the geometric mean resistance time was only 30
minutes, whereas in the static test it was found to be more than twice
as long with the same initial cyanide concentration. Even if the dif-
ference of mean resistance periods was fortuitously somewhat exaggerated,
as it well may have been in this single test, it obviously can become
very important when concentrations tested are much lower and the mean
or median resistance periods under constant-flow test conditions are as
long as 24 hours or longer, requiring correspondingly extended tests.
Were it not for the already mentioned nature of the relationship between
cyanide concentration and the survival time of fish, the results of
prolonged static tests without frequent renewal of test solutions would
have been much more misleading than they usually have been actually.
Estimates of lethal threshold concentrations derived by extrapolation
from such results obviously cannot be very reliable. By increasing
susceptibility to cyanide poisoning, the reduction of dissolved oxygen
concentrations in unaerated, standing test solutions with fish doubt-
less has tended to compensate somewhat for the loss of cyanide from
the unrenewed solutions in static tests. But how low would have been
the median tolerance limits of free cyanide determined for such very
susceptible fishes as the marine pin perch, Lagodon rhomboides, had the
constant-flow bioassay method been used and the tests continued for at
least 48 or 96 hours? The 24-hour value for the pin perch reported by
Daugherty and Garrett (1951), less than 0.07 mg/1 as CN, was determined
by a static bioassay in which the test solutions were aerated continu-
ously. Yet it is as low as, or lower than, 3-day tolerance limits and
even lethal threshold concentrations for salmonid fishes indicated by
28
-------
results of constant-flow tests under favorable conditions. A need for
more and better experiments with fishes apparently (or possibly) much
less resistant to cyanide poisoning than the salmonids is surely
indicated.
The results of numerous tests of the toxicity of free cyanide to the
bluegill have been fairly uniform. Cairns and Scheier (1958, 1959,
1963, 1968) and Patrick, Cairns, and Scheier (1968) reported 96-hour
median tolerance limits of KCN for the bluegill determined with fish of
various sizes in different dilution waters at temperatures of 18 or
20° C. These values, obtained by static bioassays with controlled
aeration of test solutions, range from about 0.17 to about 0.23 mg/1 as
CN, and they pertain to amounts of cyanide added to the waters used,
and not to analytically determined concentrations. It can be seen that
they are not very much higher than the 48-hour and 72-hour median
tolerance limits for the bluegill at 20° C (about 0.154 mg/1 as CN)
that have been determined by continuous-flow bioassay (Doudoroff, Leduc
and Schneider, 1966) and that also pertain to cyanide added (but in the
form of NaCN). It must be noted here that the 48-hour and 72-hour
median tolerance limits reported by Cairns and Scheier (1963) for the
18° C temperature evidently are incorrect because of faulty interpola-
tion. However, the experimental data needed for their correction are
fully presented in tables, and the corrected values all fall within the
range indicated above for 96-hour tolerance limits reported by the
authors. Turnbull, DeMann, and Weston (1954) found both the 24-hour and
the 48-hour median tolerance limits of KCN for the bluegill, determined
by static bioassay at 20° C, to be about 0.28 mg/1 as CN, a value
apparently much too high, especially for the 48-hour limit. On the
other hand, Broderius (1973) reported a median survival time of only
11 hours at a cyanide concentration (initial) of 0.15 mg/1 as CN in a
standing solution of NaCN at 20° C. Henderson, Pickering, and Lemke
(1961) reported a 96-hour median tolerance limit of NaCN for bluegills,
29
-------
also determined by static bioassay without renewal of test solutions at
the higher temperature of 25° C, of 0.15 rag/1 as CN. A "toxicity
threshold" of about 0.175 mg/1 as CN (0.33 mg/1 as NaCN) also has been
reported (Michigan Department of Conservation, Institute for Fisheries
Research, 1933). At the rather high temperature of 30° C, Cairns and
Scheier (1963) found 96-hour median tolerance limits of KCN for the
bluegill to be about 0.13 to 0.14 mg/1 as CN, whereas comparable values
determined in the course of the same study at 18° C were about 0.17 to
0.18 mg/1 as CN. It should be noted that none of the above tolerance
limits for the bluegill is much more than twice the lowest value
reported, and all but one are less than twice the lowest value, even
though the experimental methods and temperatures were not at all
uniform. This agreement of the bioassay results is not any better,
certainly, than the agreement to be expected among results of truly
meaningful toxicity bioassays properly performed, at least when tem-
peratures are fairly uniform. However, it compares most favorably with
the agreement, or lack of agreement, of the published results of toxicity
tests with some other species of fish used as test animals, such as the
goldfish, Carassius auratus, for which 4-day tolerance limits of free
cyanide ranging from less than 0.1 mg/1 as CN (0.04-0.12 mg/1) to more
than 1.5 mg/1 have been reported, the brown trout, etc. The vast dis-
crepancies of some of the findings reported in the published literature
are all but inexplicable; gross methodological or other errors must be
assumed in proposing plausible explanations.
RELATIONS BETWEEN CYANIDE CONCENTRATION AND SURVIVAL TIME
The relationships between free cyanide concentration and the survival
time of fishes or the overturning time, and also the relationships
between the duration of exposure to given cyanide concentrations and the
percentage of test animals surviving, have been studied and discussed by
a number of investigators (Wuhrmann and Woker, 1948; Woker and Wuhrmann,
30
-------
1950; Wuhrmann, 1952; Herbert and Merkens, 1952; Herbert and Downing,
1955; Herbert, Downing, and Merkens, 1955; Neil, 1957; Burdick, 1957;
Burdick, Dean, and Harris, 1958; Gillar, 1962; Abram, 1964; Doudoroff,
Leduc, and Schneider, 1966; Chen and Selleck, 1969). Various equations
describing the relationships mathematically have been advanced. These
are pertinent to the estimation of lethal threshold concentrations that
have been reported here, but a detailed and necessarily involved
exposition of this matter would not be appropriate to the intended
purpose of this literature review.
Some of the available data suggest a rectilinear relationship between
the logarithms of cyanide concentrations and logarithms of mean or
median resistance times over a wide range of cyanide concentrations
(Herbert and Merkens, 1952; Neil, 1957; Gillar, 1962; Doudoroff, Leduc,
and Schneider, 1966). Great deviations from this relationship at very
high, rapidly fatal cyanide concentrations have been obvious whenever
such concentrations have been tested. The existence of a minimum
(threshold) response or resistance time has been assumed by some, as
can be seen in the well-known equation:
CC - a)n CT - b) = K
where C = concentration of the toxicant,
a = threshold concentration for response (e.g., lethal
threshold or incipient lethal level),
T = duration of exposure (response or resistance time) ,
b = minimum or threshold response (or resistance) time,
and n and K are appropriate constants (Wuhrmann, 1952; Warren, 1971).
At the lowest concentrations tested, which were tolerated by rainbow
trout for more than 3 days, the data of Herbert and Merkens (1952)
revealed to them no deviation from the above-mentioned rectilinear
31
-------
relationship between the logarithms of concentration and of exposure
time (Cn T = K, where the terms are as just defined above). There can
now be no doubt, however, that lethal threshold concentrations of
cyanide do exist (i.e., are real). The existence of a threshold level
is clearly indicated even by some data that also have strongly suggested
a linear relation between logarithms of median survival times and
toxicant concentrations over wide ranges of these variables, as noted by
Doudoroff, Leduc, and Schneider (1966). One can reasonably question
only the accuracy of estimates of the threshold values that have been
variously arrived at by extrapolation from relatively short-term
lethality data, and perhaps debate the relative merits of estimation
procedures that have been employed.
Herbert and Downing (1955) made the interesting observation that the
frequency distribution of the periods of survival of individual rain-
bow trout in KCN solutions was symmetrical and approximately normal at
rapidly lethal, high concentrations of the poison only. It was
asymmetrical and approximately log-normal within that range of lower
concentrations where the relation between log concentration and log
survival time was believed to be linear, and not at concentrations well
above that range. The significance of this observation is not clear.
INFLUENCE OF TEMPERATURE
Pronounced acceleration of the lethal action of free cyanide on fish
due to increases of temperature has been observed by a number of
investigators when various fishes have been exposed to constant, more
or less rapidly lethal cyanide Concentrations at the different tempera-
tures. Linear relationships have been observed between temperature and
the reciprocal of the overturning time (Southgate, Pentelow, and
Bassindale, 1932; Alexander, Southgate, and Bassindale, 1935) or, more
often, the logarithms of the overturning or final immobilization (lethal
32
-------
exposure) time or of its reciprocal (Wuhrmann and Woker, 1953, 1955;
Sumner and Doudoroff, 1938). Nearly twofold to nearly threefold
increases of the rate of lethal action of the poison with each 10° C
rise of temperature (i.e., Q10 values of 1.8 to 2.8) have been recorded.
It should be noted, however, that these results have been obtained
always in tests at relatively high cyanide concentrations, at which
most of the test animals even of the most resistant species and tested
at the lowest temperatures were overcome within 6 hours. The fish
tested at the highest temperatures turned over or died within about an
hour or much sooner. In experiments with trout, the fish were overcome
in a few minutes at high temperatures and within less than 20 minutes,
on the average, at the lowest temperatures tested. Very different
results have been obtained by British investigators in experiments with
juvenile rainbow trout when these fish were exposed to relatively low
cyanide concentrations in the Water Pollution Research Laboratory.
Apparently, only very brief and incomplete reports of these interesting
and important results have, as yet, been published, as items included
in annual reports of progress of the Laboratory's work.
In one series of tests (Great Britain, Department of Scientific and
Industrial Research, 1953), the young rainbow trout were exposed to
different cyanide concentrations ranging from 0.125 to 1.0 mg/1 as CN
at each of three temperatures ranging from 12 to 22° C. At concentra-
tions from 1.0 to 0.3 mg/1 as CN, the expected relation between tempera-
ture and mean duration of survival, which was about 7 minutes or less,
was again seen. The fish succumbed most rapidly at the highest tempera-
ture, and the relation between temperature and log survival time at
each concentration, especially the two highest ones, deviated little
from the expected linearity, or inverse proportionality. However, at
concentrations from 0.25 to 0.175 mg/1, the temperature differences had
no appreciable effect on the mean survival periods, which were all less
than 18 minutes, and at the still lower concentrations of 0.15 and 0.125
33
-------
mg/1, the trout tended to survive longest at the highest temperature and
to succumb soonest at the lowest temperature. At the 0.15 mg/1 concen-
tration, the observed relation between log survival time and temperature
was very close to direct proportionality; the slope of the straight line
that could be fitted to the data was even greater than the slope in the
opposite direction of either straight line that could be fitted to data
obtained at the two highest concentrations tested. Even at the lowest
tested concentration, the survival time of the trout usually was less
than 200 minutes, so that not only at very slowly lethal cyanide levels
was the unexpected relation to temperature demonstrable. It was noted
that the deviation from the expected relation first became apparent at
about the same concentration as that at which the relation between the
logarithms of both cyanide concentration and mean duration of survival
began apparently to deviate from linearity as the concentration was
increased.
In the experiment just reported, no attempt was made, of course, to
estimate and compare lethal threshold concentrations of cyanide at the
different temperatures. The threshold values, however, are far more
meaningful, of course, from a practical standpoint, than average periods
of survival at rapidly lethal concentrations or lethal concentrations
pertaining to arbitrarily selected, short exposure periods. In a much
later experiment (Great Britain, Ministry of Technology, 1968), juvenile
rainbow trout 3-5 cm long and acclimated to the test temperatures for
3 to 4 days were exposed for relatively long periods in enclosed vessels
to different, continuously renewed cyanide solutions at temperatures of
12-13° C and 2-4° C. At the higher temperature, a not unusual, simple
relation was observed between cyanide concentration and median periods
of survival of the fish. A concentration tolerated by 50 percent of
the fish for about 4 days apparently was near the lethal threshold
level, which appears to have been in the neighborhood of 0.08-0.09 mg/1
as HCN. The curve relating cyanide concentration to median survival
34
-------
periods at the low temperature, on the other hand, is highly irregular,
with two sharp inflections at the survival time level of about 33 hours.
Median survival periods in the neighborhood of 33 hours were recorded at
different cyanide concentrations ranging from about 0.03 to about 0.06
mg/1 as HCN, but the survival time increased to about 4 days at a
cyanide concentration between 0.02 and 0.025 mg/1 as HCN, which appears
to have been approximately the lethal threshold level. Such an irregu-
lar relation between concentration of a toxicant and survival time is
indicative of chronic toxicity which is physiologically (i.e., with
respect to mode of action) quite distinct from rapidly lethal, acute
toxicity. The toxic action of a single substance can be very different
at high, rapidly lethal, and low, slowly lethal concentrations, and in
a mixture of toxicants, different ones can be predominantly effective
at different concentrations of the mixture. Each of two modes of action
of a toxicant, or of two toxicants in a mixture, may be characterized by
a very distinct relation between concentration and response or resist-
ance time, and an irregularity such as that in question thus is apt to
mark the transition from one mode of action to another. It should be
noted that the indicated lethal threshold concentration of cyanide
(0.02-0.025 mg/1 as HCN) at the low test temperature (2-4° C) is much
less than one-third of the corresponding value (0.08-0.09 mg/1 as HCN)
pertaining to the higher test temperature of 12-13° C. On the other
hand, extension of that portion of the curve relating cyanide concen-
tration to survival time at the low temperature which was fitted to
data obtained at concentrations not less than 0.06 mg/1 as HCN, so as
to estimate the lethal threshold level by extrapolation from these
data, would have indicated a threshold concentration of about 0.055 mg/1
as HCN. This value is much more than one-haIf the corresponding value
pertaining to the higher test temperature, which could be arrived at
without such extensive extrapolation. Extension of the two curves in
the opposite direction indicates that they would have met and crossed
each other at a point corresponding to an HCN concentration of about
35
-------
0.4 mg/1 and a median survival time of about 100 minutes, if high.
enough concentrations had been tested to define the curves to the
point of their intersection, which can be only very imprecisely esti-
mated.
In two short paragraphs, without an accompanying graph, published one
year later than the report just discussed, results of a further in-
vestigation very similar to the last-mentioned one have been reported
(Great Britain, Ministry of Technology, 1969). Rainbow trout 2.5-2.6
cm long and acclimated for 3 to 4 days to the different test tempera-
tures, which ranged from 4 to 20° C, were used. The cyanide solutions,
tested in totally enclosed vessels, were prepared with a hard water,
had pH values of 8.0 to 8.3, and were continuously renewed. It was
reported that, at high cyanide concentrations, the fish died soonest
at high temperatures, but at low concentrations at which the curves
relating survival time to concentration of poison became almost parallel
to the time axis, the fish died soonest at low temperatures. Unfortu-
nately, the concentrations at which the different relationships of
survival time to temperature were observed or at which the curves in
question intersected were not reported. It was stated that the 72-hour
median lethal concentration or tolerance limit, which is close to the
median threshold concentration for survival, increased almost threefold
with increase of temperature from 4 to 20° C, but the median lethal
concentrations determined at the different temperatures at which tests
were performed were not reported. Curiously, nothing at all was said
again about the interesting irregularity of the relation between cyanide
concentration and the median duration of survival at a very low tempera-
ture that had been previously reported. It is not clear, therefore,
whether this irregularity was again observed or was found to have been
probably attributable to error or some defect of the earlier experiment.
The indicated difference between the threshold or nearly threshold
levels of free cyanide at temperatures of 4 and 20° C is smaller than
36
-------
the difference between the corresponding values for temperatures near
3 and 12.5° C determined previously. This fact suggests that the
threshold concentration at the low temperature indicated by the initial
experiment perhaps was later found to have been erroneous (too low) for
some reason, but one can only speculate about such matters in the
absence of more complete information.
In a still later report of a related study at the same laboratory
(Great Britain, Department of the Environment, 1972}, some data on the
influence of acclimation of young rainbow trout to a low temperature
(3° C) on their resistance to cyanide poisoning at that temperature are
presented. The fish were transferred from a temperature of 10° C to
the test temperature of 3° C, and after varying periods of acclimation
to the latter temperature ranging from nil to 7 days, their median
lethal concentrations for uniform exposure periods of nearly 6, 12, and
24 hours were determined. When the periods of exposure to cyanide were
relatively short (12 hours or less), the fish that had not been accli-
mated to the low test temperature proved most resistant to the poison,
those that had been acclimated for only 1 day were next in resistance
but much less resistant, and those acclimated for longer periods were
least resistant, but there was no consistent difference in resistance
among the groups acclimated for different periods greater than 1 day.
When the duration of exposure to the cyanide solutions was 24 hours, no
relation between resistance to cyanide (median lethal concentrations)
and the duration of previous thermal acclimation was evident. The esti-
mated 24-hour median lethal concentrations or tolerance limits for the
different groups of fish were said to have averaged about 0.05 mg/1 as
HCN, but according to an accompanying graph, the average was nearer
0.04 mg/1, the individual values ranging from about 0.03 to about 0.055
mg/1 as HCN. These results of toxicity tests of fairly short duration
again reveal great sensitivity of the rainbow trout to free cyanide at
very low temperatures, suggesting that the lethal threshold level at
37
-------
3° C is less than 0.03 mg/1. The observed decrease of resistance to
cyanide that results from acclimation to the low temperature evidently
proceeds while the fish are exposed to eventually lethal levels and is
so rapid that it is virtually complete before the end of a 24-hour
exposure period.
Herbert and Merkens (1952) reported earlier that the mean duration of
survival in KCN solutions with a cyanide concentration of 0.15 mg/1 as
CN and temperature of 17.5° C of rainbow trout previously held at un-
specified, lower temperatures increased for some time as the fish
became acclimated to the warmer water. The mean exposure periods to
overturning of the fish were about 29 minutes after 1 and 2 days of
acclimation to 17.5° C, 35 minutes after 4 days, 47 minutes after 5
days, 40 minutes after 172 hours (more than 7 days), and 51 minutes
after 8 days, the longest acclimation period tried. The fish were
exposed to cyanide more than once, but this did not appear to be the
reason for their increased resistance.
It has been known for a long time that the resistance of fish to cyanide
poisoning at a given temperature can change markedly as the animals
become acclimated to that temperature after a large decrease or increase
of the water temperature. After a few preliminary experiments with the
marine longjaw goby, Gillichthys mirabilis, reported by Sumner and Wells
(1935) had shown that such changes of resistance are demonstrable, the
phenonenon was investigated in detail by Sumner and Doudoroff (1938),
who used the same marine fish as their test animal. Gobies were
acclimated for long periods to temperatures of 10, 20, and 30° C and
then transferred for varying periods to temperatures higher or lower by
10° than the original acclimation temperatures. The mean duration of
their survival in a M/1000 solution of NaCN (26 mg/1 as CN) in sea
water then was determined after each of a number of different periods of
exposure to the new acclimation temperatures. In this way, it was
38
-------
possible to demonstrate conclusively that, at any given temperature,
the fish were most resistant to the cyanide poisoning soon after their
transfer to that temperature from a higher acclimation temperature and
least resistant after transfer from a lower acclimation temperature.
The effects of previous acclimation to higher or lower temperatures
persisted for 3 days at least, and small remnants of them probably much
longer. When some fish initially conditioned to 20° C were exposed for
varying periods to the lower temperature (10° C) and others to the
higher temperature (30° C), and all were then returned to the inter-
mediate temperature for varying periods before exposure to cyanide at
that temperature, the groups of fish previously held at the low and
high temperatures differed markedly at first in their resistance. The
persistence of this effect of previous acclimation to different tempera-
tures on relative susceptibility to the poison at 20° C was found to
increase as the duration of the previous exposure to the lower and
higher temperatures increased. Thus, after acclimation to the low and
high temperatures for only a day, the effect on resistance at the
intermediate temperature persisted for a day but not for 3 days. When
the acclimation to low and high temperatures lasted for 5 days, however,
a significant difference between the mean survival periods of the groups
of fish that had been held at these different temperatures was still
demonstrable 3 days after return of the fish to the intermediate
temperature.
Large differences in resistance to cyanide (M/1000 KCN solutions) of
different populations of fishes of the genus Cyprinodon, and also of the
poeciliid genus Crenichthys, inhabiting cool and warm springs in the
Amargosa Desert of southwestern Nevada were observed by Sumner and
Sargent (1940) when comparative tests were made at widely different
temperatures to which these populations were accustomed. Although
presumably isolated for thousands of years and thoroughly adapted to
their warm-water environment, the populations inhabitating the warm
39
-------
springs were much less resistant at the high temperatures of these
springs (32.5-33 and 35.5-37° C) than were populations inhabiting cool
springs at the lower temperatures at which they were found (24 and
21° C). When tested at the same (common) temperature, low or high, the
populations of the warm springs sometimes, but not always, proved more
resistant to cyanide than were the populations of the cool springs.
It has been noted already that the 96-hour median tolerance limits of
KCN for the bluegill at the temperature of 30° C (0.13-0.14 mg/1 as CN)
that have been estimated by Cairns and Scheier (1963) are considerably
lower than the corresponding values (0.17-0.18 mg/1 as CN) for the
temperature of 18° C determined by the same authors in the course of
the same, comparative study. These results were obtained in two bio-
assays, with soft and hard water, at each temperature. The 96-hour
median tolerance limits, evaluated by static bioassays without renewal
of test solutions, cannot be regarded as true lethal threshold concen-
trations, mainly because there was a fairly rapid loss of cyanide from
the test solutions during the tests. Probably it was mainly for the
same reason that the values obtained at 18° C are somewhat higher than
the 72-hour median tolerance limit (0.154 mg/1 as CN) determined at the
temperature of 20° C by Doudoroff, Leduc, and Schneider (1966), who
employed the constant-flow technique. It should be noted, however,
that the latter tolerance limit is somewhat higher than the 96-hour
median tolerance limits (also the 72-hour values, which were identical
with the 96-hour values) at a temperature of 30° C estimated by Cairns
and Scheier. It can be concluded, therefore, that the bluegill's long-
term tolerance probably is indeed considerably greater at 18° C than it
is at the much higher temperature of 30° C. This, of course, is the
relationship that one would expect on the basis of the results of short-
term tests. Unfortunately, no tests have been performed with the blue-
gill at very low temperatures such as 3 or 4° C, or even at 10° C.
Perhaps its lethal threshold level of cyanide is maximal at a temperature
that is neither high nor low, such as 15 or 20° C.
40
-------
Hubault (1955a, 1955b) apparently found that the logarithms of the
lowest cyanide concentrations at which cyprinids, Rutilus rutilus or
Scardinius erythrophthalmus, are noticeably affected within 255-270
minutes decrease with rising temperature most markedly when the tem-
peratures are extreme, high or low. The decrease of the logarithm of
the effective concentration per degree of temperature increase seemingly
was found to be much less over a wide range of intermediate temperatures,
so that the relationship was representable by a decidedly sigmoid curve.
However, Hubault's graphical presentation of his findings is very
confusing. I was unable to determine the meaning of some unexplained
lines and notations (numbers, arrows, etc.) on his graphs (Figure 1 in
each paper), or even to determine with certainty which one of the many
curves ("isochrones") shown there pertains to cyanide. Data to which
the curves were fitted are not reported or plotted on the graphs.
Observations apparently were made on small numbers of fish. In my
opinion, the indicated relation between temperature and effective con-
centration of a toxicant is improbable.
Evidently, much still remains to be learned about the influence of
temperature on the resistance of fish to cyanide. The relatively low
resistance of rainbow trout to low, slowly lethal concentrations of
free cyanide at low temperatures that has been reported probably is a
result of decrease of the rate of detoxification at reduced tempera-
tures. When cyanide concentrations are high, the detoxification
mechanism must be soon overwhelmed by the rapid influx of the poison,
and the rate of mortification must become dependent only on the rates
of other, continuing, vital processes that constitute the disorganized,
general tissue metabolism. The speed of death then can be seen to be
directly related to the metabolic rate. Inasmuch as the fatal tissue
concentration level of a poison may decrease with increase of the
metabolic rate, and also because the rate of entry of the poison can
increase with rise of temperature (because of accelerated external
41
-------
respiration, etc.)* an increase of the detoxification rate at an elevated
temperature need not necessarily result in a higher lethal threshold con-
centration in the external medium. If, however, at a high temperature,
but not at a low temperature, a poison is eliminated from the body as
rapidly as it enters before a lethal internal concentration is reached,
the animal can continue to withstand at the high temperatures an external
concentration that is intolerable, causing death eventually, at the low
temperature. When the cyanide dose is massive, sluggish and hardy warm-
water fishes, such as the longjaw goby, G i 11 ichthys mir ab i1is, can live
very much longer at a low temperature than can active and sensitive
cold-water fishes, such as trout, exposed to high concentrations of the
poison at the same temperature. But it is reasonable to suppose that
neither the rainbow trout nor the family Salmonidae is unique in having
lower lethal threshold concentrations of cyanide at low temperatures
than at higher temperatures. Results similar to those obtained in the
experiments with rainbow trout were said to have been obtained also in
tests with the roach, Rutilus rutilus, (Great Britain, Ministry of
Technology, 1968), but supporting data were not presented. Alabaster
et al. (1972) stated that juvenile roach proved much more sensitive to
cyanide than rainbow trout in 48-hour tests at 3-4° C. The published
information on the relation to temperature of lethal threshold levels
of cyanide even for rainbow trout is incomplete, and such information
pertaining to other fishes is certainly quite inadequate. The tran-
sient influence of recent thermal experience (acclimation) of fish on
their resistance to high concentrations of cyanide at any given tempera-
ture is interesting but probably not of major practical importance.
INFLUENCE OF DISSOLVED OXYGEN
Pronounced increases of the susceptibility of fish to cyanide at reduced
concentrations of dissolved oxygen have been observed repeatedly in both
static toxicity tests (Southgate, Pentelow, and Bassindale, 1933;
Wuhrmann and Woker, 1953, 1955) and experiments with continuously
42
-------
renewed solutions (Doming, 1954; Burdick, Dean, and Harris, 1958).
Downing (1954) found the influence of dissolved oxygen on the survival
time of yearling rainbow trout (at 17° C and pH 7.8-8.2) to be greatest
at the lowest cyanide concentration tested (0.105 mg/1). At this con-
centration, there was a hundredfold increase of the median survival
time (time to immobilization) with increase of dissolved oxygen from
about 3 to about 9 mg/1. At the higher cyanide level of 0.116 mg/1,
the corresponding increase was not much more than a tenfold increase,
and at 0.155 mg/1, a level at which the median immobilization time was
less than 17 minutes even at a high oxygen concentration, the effect
was much smaller. Wuhrmann and Woker (1953, 1955), in experiments with
the minnow Phoxinus laevis at three different temperatures, also found
the effect of dissolved oxygen on resistance to cyanide (overturning
time) to be very pronounced at low levels of cyanide but not at high
levels. In Downing1s (1954) experiments, the effect of a 1 mg/1 dis-
solved oxygen increment on resistance to cyanide was not found to
decrease as the air-saturation value was approached. At low cyanide
concentrations, the effect (proportional increase of the immobilization
time) was most pronounced at the highest dissolved oxygen levels. How-
ever, the data of Wuhrmann and Woker (1953, 1955), and also those of
Southgate, Pentelow, and Bassindale (1933), who used rainbow trout as
test subjects as Downing did, but at a temperature of 7-9° C, indicate
a more pronounced effect at low dissolved oxygen levels than at high
levels.
Only Burdick, Dean, and Harris (1958) are known to have estimated
lethal threshold concentrations of cyanide at different, constant
levels of dissolved oxygen for comparative purposes. As noted already,
they estimated that reduction of the oxygen concentration (by means of
nitrogen) from a level near the air-saturation level to 5.2 mg/1 caused
a reduction of the lethal threshold concentration of cyanide for brown
trout by only about 0.01 mg/1 (from 0.09 to 0.08 mg/1 as CN) at a
temperature of about 15.5° C. Reduction of the oxygen concentration to
43
-------
4.2 mg/1 was estimated to have caused reduction of the lethal threshold
concentration of cyanide for smallmouth bass, Micropterus dolomieui, by
less than 0.02 mg/1 or about 17 percent (from 0.104 to 0.086 mg/1 as
CN) at 21° C. The mean periods of survival (to overturning) of the
trout or the bass at the high and reduced oxygen concentrations are
strikingly different (i.e. shorter at the low dissolved oxygen levels)
when the cyanide concentrations are equal and low. But they differ
relatively little when the cyanide levels are equal and high, and the
differences of the estimated lethal threshold levels of cyanide at the
different oxygen concentrations are not very impressive. Apparently,
when dissolved oxygen concentrations do not fall below 4 mg/1 or much
below 50 percent of air-saturation levels, the influence of their wide
variations on the maximum cyanide concentrations tolerated indefinitely
by fish is not of great importance. In the experiments of Burdick,
Dean, and Harris, the individual survival (overturning) times of the
fish at low cyanide and low oxygen concentrations were extremely
variable. Also, the maximum geometric mean resistance periods recorded
in the tests at reduced oxygen concentrations were rather short (202
and 362 minutes for the trout and the bass, respectively), though they
were much longer than the maximum mean or median values recorded by the
other investigators whose comparable results have been summarized above.
The indicated lethal threshold values derived by extrapolation from the
data in question are not likely to be very accurate estimates, but
they probably are not seriously misleading.
Cairns and Scheier (1958) found that the 96-hour median tolerance limit
of KCN for bluegills at 18° C was reduced to about 0.05 mg/1 as CN by
daily reduction of the dissolved oxygen content of the standing test
solutions (by means of nitrogen) to about 2 mg/1 for a period of about
2 hours. Gradual reduction of the oxygen concentration to the low level
from the air-saturation level and its subsequent return to the high level
by aeration each was accomplished in about 3 hours. In the same study,
44
-------
the 96-hour median tolerance limit for controls held at "normal"
(relatively high) oxygen concentrations of 5 to 9 mg/1 was found to be
0.18 mg/1 as CN. It appears that the occurrence of very low but not
lethal concentrations of dissolved oxygen even for limited periods can
seriously depress the cyanide tolerance of fishes.
INFLUENCE OF WATER SALINITY AND HARDNESS
Broderius (1973) found that threespine sticklebacks exposed to NaCN
solutions of equal cyanide content died much sooner in solutions pre-
pared with sea water having a chlorinity of about 16-17 parts per
thousand than in solutions prepared with half-strength, diluted sea
water or with fresh water. The pH of each test solution averaged about
7.7. At a cyanide concentration of 0.27 mg/1 as CN in fresh water, in
the 50 percent diluted sea water, and in sea water, the median survival
periods (at 20° C) were 412, 371, and 198 minutes, respectively; at a
concentration of 0.21 mg/1 as CN, they were 642, 582, and 350 minutes,
respectively. It can be seen that in the medium that was half sea
water and half fresh water, the cyanide was only a little more toxic
than it was in fresh water, so that the much greater toxicity in the
undiluted sea water could hardly have been due to any chemical inter-
action between the cyanide and the sea salts. And there was no consist-
ent relation between the salinity of the solutions and their molecular
HCN content. The determined molecular HCN concentrations in the test
solutions containing 0.21 and 0.27 mg/1 of cyanide as CN (amounts added
to the water, averaging 0.24 mg/1) that were prepared with sea water,
50 percent sea water, and fresh water averaged 0.206, 0.203, and 0.206
mg/1, respectively. These data indicate some loss of cyanide from all
the solutions but only fortuitous variation of the individual, deter-
mined levels of molecular HCN in waters of varying salinity to which the
same amount of NaCN had been added. Both the sticklebacks and the sea
water had been obtained from an estuarial source. The salinity of this
45
-------
sea water, in the neighborhood of 30 parts per thousand, was well
below that of full-strength, oceanic water (about 35 parts per thousand),
and so it could not have been an unduly saline, unfavorable medium for
the euryhaline sticklebacks, which commonly inhabit such water. The
sticklebacks, collected in salt water of unknown salinity, had been
acclimated for about a week before the tests to different waters with
the salinities of the test solutions to which they were subsequently
exposed.
According to an earlier report (Great Britain, Ministry of Technology,
1968), results of experiments with juvenile rainbow trout 4-5 cm long,
at temperatures of 11-22° C, had suggested that the toxicity of HCN is
not related to water salinity= However, the data on which the tentative
conclusion was based were not presented. The most saline water tested
was 70 percent sea water. The trout had been acclimated to .the differ-
ent water salinities. Salinity of the water may not influence the
susceptibility to free cyanide of all euryhaline fishes in the same way.
There is some published evidence that the hardness or the total alka-
linity of water per se can have a material effect on the toxicity of
free cyanide to fish, but it is not convincing; contradictory findings
pertaining to this matter have been reported. When sufficiently high
pH values are involved, pH differences associated with differences of
carbonate hardness or total alkalinity can, of course, be reasonably
expected to have some effect. Henderson, Pickering, and Lemke (1961)
did report a considerable difference of 96-hour median tolerance limits
of NaCN for fathead minnows in a soft water and a hard water at 25° C;
these estimated limits for the soft and hard waters, evaluated by static
bioassays without renewal of test solutions, were 0.23 and 0.35 mg/1 as
CN, respectively. The alkalinity and hardness of the soft water (pH 7.4)
were 16 and 20 mg/1 (as CaC03), respectively, and those of the hard
water (pH 8.2) were 320 and 380 mg/1, respectively. The pH of the well
46
-------
buffered, hard water (8.2) presumably did not increase in the presence
of respiring test animals, and such a pH is not high enough to have
greatly depressed the toxicity of the free cyanide by increasing the
ionization of HCN. Therefore, the observed difference in toxicity of
NaCN solutions prepared with the two waters appears to be attributable
to the difference in hardness or of total alkalinity. However, the dif-
ference of results obtained with the two waters may have been fortuitous.
No difference of the toxicity to fathead minnows in the two waters of
acetaldehyde cyanohydrin (lactonitrile)t whose rapid hydrolysis in
aqueous solutions yields free cyanide, was observed by the same inves-
tigators .
Cairns and Scheier (1963) observed no difference of the toxicity of KCN
to bluegills in their synthetic soft and hard waters. They estimated
the 96-hour median tolerance limits of KCN in each water at two differ-
ent temperatures, and the means of their values for the two waters are
virtually identical. However, the difference between these waters was
not ureat; the alkalinity and hardness of the soft water were a little
more than twice the corresponding values for the soft water used by
Henderson, Pickering, and Lemke, and those of the hard water were a
little less than half the corresponding values for the hard water of
the latter authors.
Leclerc and Devlaminck (1950) found KCN to be some two to three times
as toxic to minnows, Phoxinus laevi_s_, in a hard, natural water than in
distilled water at 20° C. But the duration of their static bioassays
whereby their minimum lethal levels of the poison were determined was
only 6 hours, and the pH values of the test solutions were not reported.
Perhaps the hydrolysis of KCN added to the unbuffered, distilled water
caused sufficient elevation of the initial pH of the solutions to depress
the toxicity of the free cyanide markedly for all or most of the duration
of the very brief tests.
47
-------
Burdick, Dean, and Harris (1958) found no difference of the toxicity of
KCN to brown trout in two natural waters, an acid stream water with a
total alkalinity of 6 to 12 mg/1 as CaC03 and a moderately hard spring
water with a total alkalinity of about 100 mg/1 and pH 7.5-8.3. Mean
survival periods at various concentrations of cyanide (0.145 to 2.14
mg/1 as CN) in the two waters were determined and found to fit the same
curve (relating log time to log concentration) when plotted against the
cyanide concentration. Most of the mean survival periods recorded were
within the range of 5 to 100 minutes.
EFFECTS OF ACCLIMATION TO CYANIDE
The influence of acclimation of fish to low, sublethal levels of cyanide
upon their subsequent resistance to the poison has been investigated,
but not thoroughly. Malacca (1968) found that the median survival
periods of minnows, Phoxinus phoxinus, at a cyanide (KCN) concentration
of 0.5 mg/1 as CN were increased considerably, by about 60 to 70 percent,
by previous exposure of the fish for 2 days to a lower cyanide level of
0.1 mg/1 as CN. Neil-(1957), however, reported varying results of
experiments with brook trout, Salvelinus fontinalis, which were accli-
mated for 15 days or longer to one of three sublethal concentrations of
KCN (0.01, 0.03, and 0.05 mg/1 as CN) and then exposed to different
lethal levels. The acclimated trout proved sometimes more resistant
and sometimes less resistant than the unacclimated controls, the effect
of acclimation varying with the acclimation level of cyanide and with
the lethal level tested. Fish acclimated to 0.01 mg/1 were consistently
the most resistant ones, on the average. Those previously exposed to
the higher sublethal levels (i.e., to 0.03 and 0.05 mg/1 as CN) were less
resistant; they proved less resistant even than the controls to 0.30
mg/1, but more resistant than the controls to 0.40 and 0.50 mg/1, the
resistance time varying inversely with the acclimation level. It appears
that previous exposure to some low, sublethal levels of cyanide can
increase somewhat the resistance of the trout to lethal levels by setting
48
-------
in motion some adaptive mechanism, but the adaptation can be largely or
entirely canceled or counterbalanced by concomitant injury when the
fish are exposed to higher sublethal levels. That injury evidently can
result in reduced resistance to lethal cyanide levels.
Brockway (1963) found that groups of cichlids, Cichlasoma bimaculatum,
about 4 months old and 1 g in average weight, that had been acclimated
for 40-42 days to cyanide (NaCN) concentrations of 0.02, 0.06, and 0.10
mg/1 as CN were all more resistant than unacclimated controls to not
very rapidly lethal cyanide levels. Their 24-hour median tolerance
limits at 25° C were estimated to have been 0.22, 0.24, and 0.24 mg/1
as CN, respectively, whereas the corresponding value for the controls
was 0.18 mg/1. However, when a group of fish of the same age and size
acclimated for 60-62 days to a cyanide concentration of 0.10 mg/1 as
CN were exposed to the relatively high, rapidly lethal cyanide level
of 0.42 mg/1, the acclimated fish proved far less resistant than the
unacclimated controls. The determined median periods of survival were
340 minutes for the controls but only 140 minutes for the fish previ-
ously exposed to cyanide. The significance of these results is unclear.
They are not in agreement with Neil's, whose brook trout acclimated to
relatively high but tolerable cyanide levels proved less resistant than
controls to relatively low lethal levels but more resistant than the
controls to higher lethal levels. They also do not agree with Malacca's
observations on minnows. When all of the available data on the effects
of acclimation to cyanide on lethal levels are considered together,
they are quite puzzling and indicate a need for further investigation
of this matter. All of the lethal levels of free cyanide to which fish
were exposed in the reported experiments were more or less rapidly lethal
and well above the lethal threshold concentrations. What needs to be
determined is whether the lethal threshold concentrations are or are not
increased materially by acclimation of fish to somewhat lower levels.
-------
RESISTANCE IN RELATION TO BODY SIZE AND PHYSIOLOGICAL STATE
There is no uniformity of the results of investigations also of the
relation between the size of fish of a given species and their resist-
ance to cyanide poisoning. The resistance of different species has been
reported to decrease, increase, or not change appreciably with large
increases of body size. Species of fish may indeed be unlike in this
regard, but other factors such as difference of experiments in the
manner of evaluation of relative resistance or in relationship between
the size of fish and their age may have been largely responsible for the
lack of agreement of reported experimental results.
Herbert and Merkens (1952) compared the mean periods of survival of
groups of yearling rainbow trout (i.e., fish believed to be of nearly
the same age} differing widely in mean body length in a solution of KCN
whose concentration (about 0.153 mg/1 as CN) proved quite rapidly lethal
to all of the fish, killing most of them within 40 minutes. The indi-
vidual lengths of the fish ranged from 5.5 to 17.25 cm. The larger fish
tended to succumb much sooner than the smaller ones. For example, the
mean periods of survival of thirteen fish averaging 15.0 cm in length,
ten fish averaging 9.0 cm in length, and eleven fish averaging 7.1 cm
in length were 18.4, 24.1, and 37.0 minutes, respectively. Herbert and
Downing (1955) noted a statistically significant negative correlation
between body length and overturning time of rainbow trout in KCN solu-
tions when the fish were exposed to cyanide concentrations of 0.125 and
0.150 mg/1 as CN (levels apparently then lethal to the average fish
within 1 hour or much sooner), but not when they were exposed to a higher
concentration, 0.20 mg/1 as CN. There is no apparent reason to question
the reliability of the above findings, but it would be interesting to
know how, or to what extent, the differences in resistance to cyanide
may have been related to other physiological differences that had
resulted in large differences in size of fish of the same age. Would
fish of different age but of average size for their age have differed
50
-------
correspondingly in their resistance, having become increasingly suscep-
tible as they grew larger? It has seemed reasonable to suppose that the
resistance of fish to acute cyanide poisoning, which is believed to
interfere with internal or tissue respiration, generally tends to
increase with decrease of metabolic rate (Sumner and Doudoroff, 1938).
And the metabolic rate of a juvenile fish is not likely to increase as
the fish grows larger; it tends to decrease, as a general rule. However,
it should be recalled that the resistance of rainbow trout to not very
rapidly lethal cyanide concentrations has been reported to decrease
with reduction of temperature, which certainly depresses the rate of
respiratory metabolism.
Anderson (1974), in experiments in which only mature, male guppies were
used, found a definite, direct relation between body weight and resist-
ance to KCN. A linear regression with a high correlation coefficient
i
was defined by an appropriate equation. The fish were segregated for
the tests in groups of ten, and exposed for 96 hours to continuously
renewed KCN solutions at a temperature of 25° C. The 96-hour median
lethal concentration in mg/1 for fish of a given size class was shown
to be equal to 0.147 times W°'72, where W is the weight in grams of a
group of ten fish of that size class (or the weight of an individual
fish of average size multiplied by 10). Although the value to be derived
by the use of this equation was said to be the median lethal concentra-
tion of "potassium cyanide", I am convinced, for several reasons, that
this is an error and that the concentration in question actually is, or
was meant to be, the median lethal concentration of cyanide as CN. As
noted already, other investigators have found the guppy to be highly
resistant to cyanide poisoning, their data indicating 96-hour median
tolerance limits of free cyanide much above 0.15 mg/1 as CN. Thus, even
0.147 mg/1 as CN is a surprisingly low value for the 96-hour tolerance
limit for the 0.1 g, mature males, and 0.147 mg/1 as KCN, or 0.059 mg/1
as CN, hardly could be a correct value. Free cyanide as CN doubtless
was measured when the test solutions were analyzed.
51
-------
Cairns and Scheier (1959) determined by static bioassays at 20° C the
96-hour median tolerance limits of KCN for three groups of bluegills
averaging about 3.9, 6.1, and 14.2 cm in length and about 1.0, 2.8, and
54 g, respectively, in weight. These fish doubtless were not all of
nearly the same age. The reported tolerance limits for the groups of
fish of small, medium, and large size were 0.55, 0.45, and 0.57 mg/1
as KCN (about 0.23, 0.19, and 0.24 mg/1 as CN), respectively, and it was
concluded that there was no significant difference in resistance to KCN
of the bluegills of different sizes. The experimental material was
obtained from two different sources and it is not clear whether or not
this could have materially affected the result of the experiment, since
the relative numbers of fish from the two sources and their size distri-
butions were not reported. Also, although the test solutions were said
to have been analyzed for cyanide to "make certain that a rather con-
stant concentration of cyanide ions was maintained throughout the experi-
ment," the results of these analyses and of determinations made of
dissolved oxygen were not reported. It is unlikely that the cyanide
concentrations in the gently aerated test solutions remained quite
constant for 96 hours and that dissolved oxygen concentrations in all
test vessels (each containing ten small, ten medium, or five large blue-
gills) were identical. Since the fish of different sizes were not
exposed together to the same solutions, even moderate differences among
initially like solutions to which they were exposed could have materially
influenced the test results, perhaps concealing considerable differences
in cyanide tolerance of fish differing greatly in size.
Sumner and Doudoroff (1948) found no significant difference in resist-
ance to a very high concentration of KCN (M/1000) between small longjaw
gobies, Gillichthys mirabilis, and large ones whose average weight was
four to five times that of the small ones. Wells (1916), however, has
presented data showing that the average survival periods of rock bass,
Ambloplites rupestris, exposed to a concentration of KCN (M/25,000, or
52
-------
1.0 mg/1 as CN) that was also quite rapidly fatal increased markedly
with large increases of body size and, doubtless, of the age of the
fish. In his experiment on effects of starvation on cyanide tolerance,
survival periods of unstarved controls weighing 1-1.5, 10-15, 25-40,
and 80-200 g averaged 71, 106, 141, and 166 minutes, respectively. The
corresponding values for his experimental fish starved for varying
periods are 87, 107, 142, and 170 minutes, respectively. There was no
consistent difference in resistance shown between the smaller and larger
fish in the 80 to 200 g weight range, the average survival periods of
fish weighing 130-200 g having been not much greater than those of fish
weighing 80-95 g, and those of fish weighing 100-125 g having been much
less. The number of fish tested was small. Variations of the resist-
ance of fish to cyanide with body size thus may be often, but not always,
readily demonstrable, and the resistance apparently can either increase
or decrease with increase in size.
Wells (1916) believed that he was able to demonstrate effects of star-
vation on the resistance of rock bass to cyanide, which he thought
initially increased and then decreased with increasing duration of star-
vation. Although there may be such effects, my own analysis of Wells'
data led me to the conclusion that no effect actually has been demon-
strated by the few experiments performed by him. After 47 days of
starvation, the rock bass did survive longer in a KCN solution (1.0
mg/1 as CN) than controls did in five of six trials with fish of dif-
ferent size, but this result could have been fortuitous. And in the
other tests, with fish starved for 12 and 52 days, no effect or consist-
ent difference between the experimental and control fish was apparent to
me.
Costa (I966) found that minnows, Phoxinus phoxinus, that had been exer-
cised before their exposure to NaCN solutions lost their equilibrium in
very rapidly lethal cyanide solutions sooner than did unexercised
53
-------
controls. At lower cyanide concentrations that were still highly toxic,
causing loss of equilibrium after about 1 hour of exposure, there was
little or no difference, however, between the responses of the exercised
fish and the controls.
Pronounced differences in resistance to cyanide between lots of fish of
the same species and nearly the same size but obtained from different
sources and/or tested at different times have been recorded. The
reasons for these differences are unknown. Such differences in resist-
ance of young rainbow or brown trout, due perhaps to inherent physio-
logical differences between the stocks, to seasonal variations in
physiological state, or to differences in thermal history, have been
reported by British investigators (Great Britain, Department of Scien-
tific and Industrial Research, 1953; Herbert and Downing, 1955) and by
Burdick, Dean, and Harris (1958). Herbert and Merkens (1952) and Herbert
and Downing (1955) found a strong correlation between the overturning
times in separate trials of individual rainbow trout exposed to KCN
solutions and returned to clean water for recovery to permit repetition
of the relative resistance tests with the same fish. Differences in
size of the fish and the observed relation between resistance to cyanide
and body size must have been partly responsible for the above correla-
tion. Effects of thermal acclimation, as well as of previous exposure
to sublethal levels of cyanide, on the resistance of fish to the poison,
already have been discussed. Heritable (genetic) and inherent seasonal
differences of resistance have not been studied, but, as suggested
above, may well be major reasons for variability of experimental results
obtained by the most refined methods.
INTERACTIONS OF FREE CYANIDE WITH OTHER POISONS
The acute toxicity to fish of specially prepared mixtures of toxicants
including free cyanide as one of the components has been studied by
54
-------
several investigators. The toxicity of various, individual metal-
cyanide complexes will be fully considered in another section of this
review (Section V). There, the interaction of free cyanide and individ-
ual heavy metal cations, such as zinc ion, which combine to form these
complexes or derive from their dissociation, also will be discussed.
Here, only interactions of cyanide with toxicants other than the heavy
metals, or with metals present in mixtures together with other toxicants,
will be considered.
Southgate (1932) reported that marked increases of the toxicity to rain-
bow trout of rather rapidly lethal solutions of p-cresol resulted from
the addition to them of sublethal or not nearly as rapidly lethal amounts
of KCN. No corresponding increases, except a small increase in one
instance, of the toxicity of rapidly lethal KCN solutions (reported as
the reciprocal of overturning time) resulted from similar additions of
p-cresol in sublethal or much less rapidly lethal amounts. Thus,
although some infra-additive, or less-than-additive, interaction (Warren,
1971; Sprague, 1970) of the two toxicants was revealed, their joint
action certainly was not shown to be nearly additive. Because of the
high (rapidly lethal) concentration of one or the other toxicant in each
of the mixtures, all of which caused overturning of the fish of average
resistance within 25 minutes or less, the practical significance of
these observations is uncertain.
Pronounced synergism of free cyanide and ammonia (molecular base),
observed in experiments with the chub (cyprinid) Squalius cephalus at
12-14° C, has been reported by Wuhrmann and Woker (1948). Again the
tests were of rather short duration, but some tests of solutions with
concentrations of both toxicants apparently near or below lethal thresh-
old levels of the individual toxicants were included in the small series
of trials of different combinations. Curiously, when free cyanide con-
centrations were very high (7.6-15.2 mg/1 as CN) and not by themselves
55
-------
much, more rapidly lethal than were much lower concentrations, the lethal
action was accelerated in the presence of small, relatively harmless
amounts of molecular ammonia (NHg) much more markedly than it was at the
lower levels of cyanide. Indeed, when the cyanide concentration was
about 2.3 mg/1 as CN, the presence also of molecular ammonia had no
effect when its level was 0.28 mg/1 and only a moderate effect when the
level was 0.65 mg/1; yet, only 0.55 mg/1 had a striking effect when the
cyanide concentration was more than 15 mg/1, for the rate of intoxica-
tion then increased much more than threefold. Of more interest from a
practical standpoint, however, were two tests in which cyanide concen-
trations of 0.10 and 0.14 mg/1. as CN were combined with NHg concentra-
tions of 0.70 and 0.35 mg/1, respectively, and the two mixtures found
to be decidedly toxic, the fish having turned over in only 78 and 225
minutes, on the average, and died in 156 and 300 minutes, respectively.
In the absence of the cyanide, an NH^ concentration of 0.70 mg/1 did
not cause overturning of the fish within an observation period of 500
minutes or more, and 1.3 (1.2-1.5) mg/1 caused overturning only after
about 215 minutes, on the average. Cyanide alone did not cause over-
turning or death of Squalius within the observation period at a concen-
tration of 0.20 mg/1 as CN, and a concentration of 0.50 mg/1 caused
overturning only after 141 minutes and death after 268 minutes of
exposure, on the average. It is obvious that the toxicity of the mix-
tures was far greater than that attributable to independent action of
either one of the two components known to be highly toxic, but the
design of the experiments was not such as to permit complete determina-
tion of the nature and degree of interaction.
Lloyd and Jordan (1963, 1964) compared the toxicities to rainbow trout
(i.e., the median tolerance limits, or toxicity indices derived there-
from) of various sewage effluents containing cyanides with their
"predicted" toxicities. The predicted toxicities were computed values
based on chemical analyses of the'effluents, available data on the
56
-------
individual toxicities of their known toxic components Ctested separate-
ly) , and the assumption that the toxicity of each of these mixtures of
poisons is equal to the sum of the individual toxicities of the identi-
fied and measured toxic components. In other words, recognition and
appropriate measurement of all the important toxic components of the
effluents, and also strictly additive joint action of these components
(also known as additive synergism) were assumed. Fair agreement of the
observed toxicities of the effluents and those predicted by the summa-
tion of the toxicities of their known toxic constituents (expressed in
toxicity units) was found in many but not all instances. Cyanide appar-
ently was not one of the major or important constituents of some of the
effluents, though present in measurable amounts. The authors realized
that some of the cyanide and certain heavy metals, such as copper and
nickel, could have been present in the effluents in the form of rela-
tively harmless metal-cyanide complexes (complex anions) and that their
measurements of "free cyanide" were not reliable. They recognized the
possibility that disagreement of observed and predicted toxicities in
some instances could well have been due to this source of error.
Similar studies of heavily polluted river and estuarine waters have
been reported by Brown, Shurben, and Shaw (1970). These authors, how-
ever, used a new and apparently fairly reliable method for determination
of free, molecular HCN. They and Lloyd and Jordan have underestimated,
sometimes rather seriously, the toxicities of many of the polluted
waters or effluents examined; in other words, the observed toxicities
often have been considerably greater than the predicted ones. Since
the reasons for disagreement of the observed and predicted toxicities
of such complex mixtures of incompletely known composition can never be
all definitely established, studies like those just described cannot, of
course, throw much light on the interactions of cyanide with other
poisons in the mixtures. Many more experiments with various mixtures of
known and adjustable composition obviously are needed.
57
-------
It has been reported (Great Britain, Ministry of Technology, 1969) that,
when cyanide and phenol at concentrations of equal toxicity to rainbow
trout were combined in water with salinity less than 20 percent of the
salinity of sea water, the toxicity of the mixture to the trout was near
that which could be predicted on the basis of the assumption of strictly
additive interaction of the two toxicants. When, however, the water
salinities were between 50 and 70 percent of the salinity of sea water,
the observed 48-hour median lethal concentration of the mixture was
about 1.7 times the value so predicted.
Anderson (1974) evaluated the toxicity to mature, male guppies of mix-
tures of the cyanide and the pentachlorphenate of potassium, as well as
of each of the individual toxicants, in continuously renewed solutions
tested for 96-hour exposure periods. He found that the toxicity of the
mixtures was less than that predictable on the basis of the assumption
of what he termed "concentration additive" interaction of the toxicants
(i.e., the kind of interaction assumed by Lloyd and Jordan and by Brown,
Shurben, and Shaw), even though the concentration-response curves
(regression lines relating percent mortality probits to log concentra-
tion, with appropriate correction for mean body weight differences) for
the two individual toxicants proved similar in slope. So-called
"response addition", or "independent joint action" of the two toxicants,
rather than "concentration addition" ("toxic unit summation" or "similar
joint action" of the poisons), was suggested by Anderson's data but was
not definitely established. For an explanation of these terms, adopted
by Anderson, and of related, interesting concepts, the reader must be
referred to Anderson's original work. Only a very brief explanation of
the terms can be given here.
When two quite independently acting toxicants occur together, each in
concentration sufficient to kill some (e.g., 50%) but not all fish in a
given exposure period, some of a group of fish exposed to the mixture
58
-------
that would not be killed in that period by one of the toxicants may be
killed by the other. The result is one kind of "response addition".
True "concentration addition" occurs when two toxicants in a mixture
Cthe concentrations of both of which may be below their individual
effective or lethal levels) act together in effecting a response, such
as death, as though they were one and the same toxicant, except that
their equally toxic concentrations may be quite different; their rela-
tive concentrations in a mixture to whose toxicity they contribute
equally must be inversely proportional to their individual toxicities.
I must point out here that I believe Anderson erred when he stated, in
his footnote on page 54, that the equation given for derivation of the
proportion of individuals responding to a mixture when there is only
response addition involves the assumption of "a total positive correla-
tion of tolerances by the test animals to each of the toxic constituents."
On the contrary, an unlikely total lack of such correlation must be
assumed, I believe, if the equation in question is to be accepted as a
valid or appropriate representation of response addition. A true lack
of any interaction or joint action of two toxicants, a hypothetical
situation perhaps never fully realized but postulated by Warren (1971),
for example, requires a total positive correlation of relative, individ-
ual susceptibilities of the test animals to the two toxicants and,
therefore, the absence of any response addition.
Cairns and Scheier (1968) were unable to demonstrate any joint toxic
action or synergism of KCN, naphthenic acids, and potassium dichromate,
K2Cr2°7 (th-e last Present only in amounts far below lethal levels)
when these toxicants were combined in solutions and the 96-hour median
tolerance limits of the mixtures for bluegills determined. The result
of one such bioassay of a solution in which the relative amount of
cyanide was quite small, so that it could not have contributed much, to
the toxicity of the mixture, did suggest the possibility of an inter-
action of this nature. However, in another, similar test of a mixture
59
-------
in which the relative amount of cyanide was much greater, some possible
antagonism of the combined toxicants, rather than any kind of synergism,
was indicated. The reported 96-hour median tolerance limit of the mix-
ture was a level at which the concentration of naphthenic acids alone
was sufficient, and that of the cyanide (0.24 mg/1 as CN) apparently
much more than sufficient, to kill 50 percent of the test animals in
96 hours. The 96-hour tolerance limits of the individual toxicants and
of the mixtures were determined for comparative purposes, at 18° C, by
static bioassay with mild, controlled aeration and no renewal of test
solutions, and these test results (estimates) cannot be deemed entirely
reliable.
Brockway (1963) found that juvenile cichlids, Cichlasoma bimaculatum,
that had been exposed for 60-62 days to a sublethal concentration of
sodium pentachlorphenate (0.02 mg/1' as pentachldrphenol) were less
resistant than controls were to a high, lethal concentration of free
cyanide (0.42 mg/1 as CN), to which they were subsequently exposed.
Their median survival time in the cyanide solution was 253 minutes,
and that of controls was 340 minutes. Resistance to a lethal level of
pentachlorphenate was not reduced, but cichlids that had been exposed
for 60-62 days to a high but sublethal level of free cyanide (0.10 mg/1)
proved slightly less resistant than controls were*to: the lethal concen-
tration of sodium pentachlorphenate (1.1 mg/1 as pentachlorphenol).
ANTAGONISTIC ACTION OF THIOSULFATE
Under certain conditions, fish have been reported to have been evidently
protected against the lethal action of free cyanide by the presence in
the cyanide solutions of the thiosulfate (hyposufite) ion, S203"2, in
relatively high concentrations (Achard and Binet, 1934; Costa, 1965) or
by previous exposure of the fish (young carp, Cyprinus carpio) to solu-
tions of this antidote (Achard and Binet, 1934). In experiments in
60
-------
which the concentration of sodium thiosulfate, Na2S203, was constant
C0.025N, or 1,975 mg/1), Costa (1965) found the antidote to be most
effective at low cyanide concentrations, affording little or no protec-
tion against very rapidly lethal cyanide levels. None of the effects on
survival time that he observed in experiments with several species of
fish, such as a mere doubling of the survival time at a given cyanide
concentration, are very impressive, but more striking positive results
perhaps could have been obtained by performing tests at more slowly
lethal levels of cyanide. Weiss, Abramson, and Baron (1958) reported
that Na2S203 did not block the lethal action of cyanide in the Siamese
fighting fish, Betta splendens, under the particular conditions of
their experiments.
The reported antagonistic or antidotal effect of thiosulfate is believed
to be directly related to its role in the internal detoxification of
cyanide, an enzyme-catalyzed process or reaction whereby the cyanide
radical is combined in animal tissues with sulfur, the thiocyanate so
produced being relatively harmless. This antidotal effect is of some
physiological interest, but in view of the high thiosulfate concentra-
tions in the external medium that are apparently required for effective
protection of fish, 'it can hardly be reasonably regarded as having any
practical significance or value in connection with waste disposal
problems. The pertinent data will not, therefore, be considered here .
in detail. •
FIELD OBSERVATIONS OF CYANIDE-CAUSED FISH MORTALITIES
Some observations made in the field on fish mortalities in waters pol-
luted with .cyanide have been in fair agreement with the results of
tpxicity tests performed in the laboratory. Grau and Hrubec (1965),
for example, reported that a September fish mortality in a polluted
river ended.at a point downstream, from the source of temporary pollution
61
-------
where the determined cyanide concentration had declined to 0.14 mg/1 as
CM. However, some of the cyanide measured may not have been in the
free state. In reporting results of another, similar study, Moore and
Kin (1968) stated that "analytical tests on the receiving stream and
observations made on fish in the stream and in live boxes show that the
free cyanide was lethal to fish at concentrations above 0.1 mg/1, was
lethal to most game fish at concentrations of 0.05 to 0.1 mg/1, and
some fish at concentrations as low as 0.03 mg/1." These field observa-
tions were made in winter, when water temperatures were near the
freezing point. The low temperature may have been responsible for the
reported, unusually high sensitivity of some of the warm-water fish to
the free cyanide, which derived from hydrolytic decomposition of spilled
acetone cyanohydrin, (CH^^C^GHOCN. However, the data on which the con-
clusions of Moore and Kin were based were not reported in full detail,
and it is not obvious that the observed fish mortality could be cor-
rectly attributed to the measured cyanide alone. There is no apparent
reason to believe that serious analytical errors were involved, but
several chemicals other than cyanide also were or may have been involved.
Free chlorine was introduced intentionally into the stream water by
addition of calcium hypochlorite to reduce the concentration of cyanide,
converting it to cyanate, and chlorine is known first to combine with
free cyanide so as to form very toxic cyanogen chloride, CMCI. The
pollution incident investigated evidently was not a simple one. Indeed,
the authors themselves stated that the "characteristic odor of cyanide
in the stream water gave a clue to at least one cause of the kill"
(emphasis added), thus suggesting that they were not sure that free
cyanide was the only harmful agent present.
Bassindale, Southgate, and Pentelow (1933) reported some observations
made in the field on the color of the gills of Atlantic salmon, Salmo
salar, and anadromous brown trout ("sea trout") smolts that were still
alive but dying in cyanide-polluted water of the estuary of the River
Tees. They found the gill color of the dying fish, measured on a color
62
-------
scale by comparison with specially prepared color standards arranged in
a graded color chart, was a decidedly brighter red color, on the average,
than that of the gills of normal fish. A similar brightening of the red
color of the gills was observed also in brown (sea) trout and rainbow
trout exposed to lethal cyanide solutions in the laboratory, and it Was
attributed to interference by cyanide with tissue respiration and the
consequent, abnormally high degree of oxygenation of venous blood. On
the other hand, when the fish were poisoned with phenolic substances or
naphthalene or were dying of dissolved oxygen deficiency, a darkening
of the gill color was observed. The field observations on the gill
color of dying fish supported the conclusion, based on chemical analyses
of the water, that cyanide poisoning was the cause of death. In the
laboratory, the bright red color of the gills of salmonid fishes poisoned
with cyanide has been observed also by Karsten (1934) t who attributed it
to the formation of "cyano-hemoglobin", and by other investigators.
63
-------
SECTION IV
SUBLETHAL TOXICITY OF FREE CYANIDE
AND AVOIDANCE REACTIONS OF FISH
EFFECTS ON SWIMMING ABILITY
Neil (1957) found that the swimming ability of brook trout, Salvelinus
fontinalis, was impaired very markedly by a month-long exposure to 0.01
mg/1 of free cyanide, and even more at concentrations of 0.03 and 0.05
mg/1. The average duration of swimming at a given speed in a rotating,
annular chamber (at 9° C) was reduced by about 75, 90, and 98 percent,
respectively, at these levels; it was reduced by about 65 and 95 percent
after exposure to the highest concentration (0.05 mg/1) for only 21
minutes and 1 day, respectively. Trout exposed to the latter concentra-
tion for 35 or 40 days in a continuously renewed medium (KCN solution)
showed some improvement of swimming performance at once after return to
clean water, but not much additional improvement was apparent after 4
days, and the mean swimming time was about 80 percent of that of
controls (which averaged about 25-26 minutes) even when the cyanide-
exposed fish had been held in uncontaminated water for 24 (or 20?) days.
Broderius (1970) confirmed these findings in experimenting with young
coho salmon, Oncorhynchus kisutch at 15° C. He exposed the fish to the
same three cyanide concentrations in continuously renewed NaCN solutions
for 2 to 194 hours. He then determined the average duration of their
swimming in a tubular chamber against a current of high velocity
64
-------
(which was resisted by controls for about 8.7 minutes, on the average)
both in the presence of cyanide and after return of the fish to clean
water for varying periods ranging from 6 to 337 hours. After only 2
hours of exposure to cyanide, the impairment of the swimming perform-
ance of the fish was nearly or quite maximal, changing little there-
after until the exposure was discontinued. The reductions of swimming
time were not quite as great as those reported by Neil for the brook
trout but were nevertheless striking, ranging from 56 percent Cat 0.01
mg/1 CN) to 84 percent (at 0.05 mg/1 CN) after only 2 hours of exposure
to the cyanide solutions. Recovery of the swimming ability of fish
exposed to the cyanide solutions for 193 hours before return to clean
water was slow after some initial improvement, which was shown mostly
by the fish exposed to the two higher concentrations. Fish that had
been exposed to 0.01 mg/1 CN showed little improvement after 251 hours.
Even after 337 hours some impairment of swimming -ability was shown by
the fish from all the tested cyanide concentrations, especially the two
higher ones. Unfortunately, no tests were performed by either Broderius
or Neil at cyanide concentrations below 0.01 mg/1 as CN, so that the
minimum decidedly effective concentration is unknown.
Leduc (1966) studied the influence of cyanide (NaCN) on the swimming
performance of juvenile cichlids, Cichlasoma bimaeulatum (L), at 25° C.
The fish were exposed to the tested levels of free cyanide for about
one month before the swimming performance tests and during these tests.
The observed effects of low cyanide concentrations on the swimming
ability of the cichlid were not nearly as great as the effects observed
in the experiments with salmonid fishes reported above. Even at a free
cyanide concentration of 0.04 mg/1 as HCN, there was no evident effect
or only a moderate effect (a reduction by about 30 percent) on the
average duration of swimming against currents of four constant, high
velocities, which were resisted by controls for about 2, 4.5, 13, and
26 minutes, on the average. The maximum swimming speed sustained in
65
-------
tests in which, the current velocity was raised by degrees at 10-minute
intervals apparently was reduced by- less than 10 percent, on the
average, at this concentration and was not reduced appreciably at a
concentration of 0.02 mg/1 as HCN. The effects of these concentra-
tions on the average swimming time at the constant water velocities
decreased markedly as the velocities increased, no effect being appar-
ent at the highest velocities tested. At free cyanide concentrations
of 0.09 to 0.10 mg/1 as HCN, the effects on swimming ability were more
pronounced, but the swimming times at the four constant, high veloc-
ities tested were reduced by not much more than 50 percent, on the
average (45-67%), and the maximum sustained swimming speeds by only
about 25 percent. These concentrations are not far below levels that
are lethal for the young cichlids at the experimental temperature of
25° C, the 48-hour median tolerance limit being about 0.14 mg/1 as
HCN CBrockway, 1963).
In preliminary experiments with the same fish (juvenile cichlids)
exposed to cyanide for 30 days before the swimming performance tests
but not during these tests, Brockway (1963) obtained results similar
to Leduc's. The duration of swimming against a current of moderate
(lowest tested), constant velocity was markedly reduced by previous
exposure to cyanide concentrations of 0.056 and 0.10 mg/1 as CN,
but not 0.02 mg/1, as compared with that of controls. Only exposure
to the highest concentration tested (0.10 mg/1) had an appreciable
effect on the duration of swimming at three higher velocities.
The observed effects of sublethal cyanide concentrations on the swim-
ming performance of fish, found in the experiments with salmonids to
be surprisingly lasting, may not be disregarded, but the ecological
significance of the findings is uncertain. Fish often must swim as
fast as they can for a very short time in escaping their enemies or
66
-------
pursuing their prey, hut most of them may never need to do so for long
periods, even periods as long as a minute. It has been noted that the
maximum duration of very- rapid swimming of the cichlids was not appre-
ciably affected at moderate cyanide concentrations far below lethal
levels. And the large reductions of the maximum duration of swimming
at the high and lower velocities of cichlids exposed to relatively high
cyanide concentrations were accompanied by only moderate (less pro-
nounced) reductions of the maximum swimming speed sustainable for 10
minutes. Effects of cyanide on the different kinds or indices of
swimming ability of salmonid fishes, whose ability to swim at a very
high speed and whose maximum sustained swimming speeds after exposure
to cyanide were not evaluated, may well be similarly related. It is
not obvious or necessarily true that a fish whose ability to swim for
a long time at a speed well below its maximal swimming speed, but not
for short periods at nearly maximal speeds, has been somewhat impaired
is at a serious disadvantage in its natural environment.
EFFECTS ON GROWTH, FOOD CONSUMPTION, AND FOOD UTILIZATION
Leduc (1966) studied the influence of free cyanide (0.008-0k10 mg/1 as
HCN) also on the growth, food consumption, and food conversion efficiency
of the juvenile cichlids, Cichlasoma bimaculatum, which were held in
troughs with continuously renewed NaCN solutions at 25° C and with
unlimited supplies of live food (tubificid worms). He found that at
free cyanide concentrations above 0.06 mg/1 and as high as 0.09-0.10
mg/1 as HCN, the growth rate was markedly depressed, as compared with
that of controls, during the first 12 days of his 36-day experiments.
However, the initially pronounced adverse effect on growth was less
pronounced during the next 12 days of the experiments, and the growth
was faster than that of the controls during the final 12 days. Conse-
quently, there was but little effect on the over-all weight gain in the
course of the entire 36-day experimental period, and it appears that,
67
-------
had the experiments been continued for a longer period, no adverse
effect at all probably would have been revealed by the total weight
gains recorded at the end of that period. At much lower cyanide con-
centrations, growth tended to be somewhat faster than that of controls
at the beginning of the experiments, but slower than that of controls
during the final 12 days of the tests. The over-all weight gains of
the experimental fish at these concentrations during the entire 36-day
tests were not consistently smaller or greater than those of controls,
neither retardation nor stimulation of growth having been a predominant
effect of the cyanide. The observed change of the response to the high
concentrations suggests acclimation to cyanide, but the indicated,
quite different responses to very low concentrations are not readily
explainable.
At the high cyanide concentrations, the efficiency of food conversion
was consistently reduced at all times, whereas the rate of food con-
sumption was usually higher than that of controls held in uncontaminated
water, and this difference tended to increase as the experiment pro-
gressed. The increase of food consumption effectively compensated for
the marked impairment of conversion efficiency, and it must have been
the reason for the improvement of growth seen after the fish had been
exposed to the cyanide solutions for some time. At low cyanide concen-
trations, food consumption curiously tended to be reduced slightly
during the last 12 days of the experiments, but the amounts of food
consumed during the entire experimental periods tended to be slightly
greater than those consumed by controls. Each food conversion ratio
(weight gain / weight of food consumed) for the entire experimental
period was slightly less than that determined for the controls.
Under natural conditions, in the absence of an unlimited supply of food
that can be obtained with little or no effort, an increase of food con-
sumption compensating for a reduction of food conversion efficiency is
68
-------
not usually possible. Reduction of activity in the presence of cyanide
may even result in a reduction of the amount of food that can be con-
sumed. The need for experiments in which the food supplies are uni-
formly restricted in evaluating effects of cyanide on growth, and
especially for experiments in which feeding is more natural (less
effortless) than it is in small aquaria stocked with an abundance of
easily found and captured food organisms, thus is clearly indicated.
In a single ;24-day experiment .with juvenile coho salmon; Oncorhynchus
kisutch, at 16° C, Leduc (1966) found that salmon exposed in large
bottles to free cyanide (NaCN) concentrations of 0.02 to 0.08 mg/1 as
HCN grew somewhat faster than controls did during the second half of
the experiment. Only at the highest one of these concentrations, and
not at 0.02 and 0.04 mg/1, was the mean weight gain after 24 days
considerably less than that of the controls, because of>initial impair-
ment of growth, which was noted also but was less pronounced at the
lower concentrations. Adaptation to cyanide is indicated by the
improvement of growth after the initial impairment. The food conver-
sion ratio and not the food intake became higher than that of the
controls. As in the experiments with cichlids, the test solutions were
renewed continuously and the food supply (earthworms) was unlimited.
The higher test concentrations of free cyanide (0.04 and 0.08 mg/1
as HCN) are not very far below the lethal level (about 0.10 mg/1 as HCN)
for juvenile coho salmon, and they are about 4 to 8 times as great as
a level (0.01 mg/1 as CN) that has been shown to have a dramatic
effect on the swimming performance of these fish (Broderius, 1970).
It appears, therefore, that growth rate is not a very sensitive indi-
cator or measure of cyanide poisoning of juvenile salmonids, as well
as of the somewhat more tolerant cichlids tested by Leduc. The swim-
ming performance of the salmonids is clearly a more sensitive indicator
or measure, but, as noted already, it is not clear how important rela-
tion to the success of the animals in their natural environment are the
69
-------
observed effects of low cyanide concentrations on their swimming ability.
Their spontaneous and feeding activities may well be depressed also.
In two of Leduc's experiments with the cichlids, Cichlasoma bimaculatum,
in which the growth of the fish was most rapid, body proportions were
found to have been altered by prolonged exposure of the animals to the
higher cyanide levels tested. An excessive increase of the relative
depth and width of the body and caudal peduncle suggested impairment
of growth in length. Also, the fins of the fish that had been exposed
to the highest tested free cyanide concentrations (0.09-0*10 mg/1 as
HCN) were found to be abnormally brittle. No pronounced effect of
exposure to cyanide on the fat content of the bodies of the fish was
apparent.
EFFECTS ON EMBRYONIC DEVELOPMENT, RESPIRATION, AND HEART BEAT
Several physiological studies having to do with effects of sublethal
cyanide poisoning on fish embryos, but not directed toward the deter-
mination of maximum concentrations of free cyanide that are entirely
harmless to the embryos, have been reported a long time ago. The
already mentioned resistance of fish embryos to relatively high concen-
trations of free cyanide may be one reason for the apparent dearth of
published information concerning limits of contamination of water with
this toxicant compatible with successful embryonic development and
hatching. More information on this subject is clearly needed.
Philips (1940) found that the rate of oxygen consumption by embryos of
the mummichog, Fundulus heteroclitus, exposed to M/1000 NaCN solutions
in sea water (26 mg/1 as CN) was maximally reduced within 1 hour. The
remaining respiration, which amounted, during the first 6 hours after
fertilization of the eggs, to about 32 percent of the average normal
respiration, was considered to be a cyanide-stable or cyanide-insensitive
70
-------
portion of the normal respiration. The absolute value of this portion
of the normal respiratory rate increased only slightly during the first
day after fertilization and remained nearly constant thereafter for 3
additional days of observation. During this time, the total, normal
respiration increased greatly as development proceeded, this increase
obviously having been due to a progressive increase of the cyanide-
sensitive portion. Development of the poisoned embryos did not cease,
however, as soon as the respiratory rate was reduced to the minimal
level. Embryos whose exposure to the highest tested cyanide concen-
tration CM/1000) began soon after egg fertilization attained late
blastula stages or the beginning of gastrulation before the development
was completely inhibited, and the rate of their development was not much
slower than that of controls. At lower concentrations of cyanide,
development of surviving embryos proceeded to later embryonic stages;
at M/8,000 and M/16,000 concentrations (3.25 and 1.6 mg/1 as CN) it
continued after more than 2 days of exposure, but became slower than
that of controls. Thus, mummichog embryos proved capable of extensive
development, at least before the end of gastrulation, at high concentra-
tions of free cyanide. In M/2000 solutions, embryos whose exposure to
the cyanide was postponed continued developing for a shorter time but
attained a more advanced developmental stage before development ceased
than did those first exposed soon after egg fertilization. Embryos
whose development had been arrested were able to continue developing
after return to clean sea water. The pH of all the experimental NaCN
solutions used by Philips was adjusted to that of normal sea water by
the addition of hydrochloric acid (HC1), and experimental temperatures
were 20-25° C.
Pelagic eggs (embryos) of marine fishes such as the cunner, Tautogolabrus
adspersus, were found by Philips to be more sensitive to free cyanide
than were those of the mummichog. At a concentration of M/10,000 (2.6
mg/1 as CN), development of cunner embryos proceeded almost not at all
71
-------
beyond the initial, two-cell stage, and even at the concentration of
M/40,000 (0.65 rag/1 as CN) attainment of early developmental stages was
delayed, and the development was arrested in some 4-5 hours, when an
early high blastula stage had been reached. Complete inhibition of
development at these high concentrations was followed within a few hours
by disintegration of the embryos. At the lower concentrations of
M/80,000 and M/100,000 CO.325 and 0.26 mg/1 as CN), early development
was not retarded and later development was not arrested during the 27-
hour observation period. However, there was delay of attainment of
advanced developmental stages. Apparently, development in the M/80,000
solution had been retarded by about 7 hours by the end of the observa-
tion period, when the embryos in that solution extended about half way
around the yolk. At the same time, the embryos in the M/100,000 solu-
tion and in normal sea water extended about two-thirds and five-sixths
of the distance around the yolk, respectively. Even the lowest of the
tested concentrations is not, of course, a very low one, as compared
with levels lethal to sensitive, fully developed fishes.
Fisher and Ohnell (1940) observed that the frequency of beating of the
hearts of mummichog, Pundulus heteroclitus, embryos exposed to M/1000
and more dilute NaCN solutions decreased within about 2 hours to dif-
ferent, constant levels. They considered only the lowest constant
frequency attainable by increasing the cyanide concentration as repre-
senting a truly cyanide-stable or cyanide-insensitive portion of the
over-all, normal frequency. These authors found 72 percent of the
normal heart beat frequency of mummichog embryos of unspecified age to
be cyanide-sensitive, and 28 percent to be cyanide-stable, being
virtually unaffected even by a fourfold increase of cyanide concentra-
tion beyond M/1000 (26 mg/1 as CN), the minimum level that was said to
be necessary to reduce the frequency by 72 percent. The test solutions
were prepared by diluting with distilled water a solution of NaCN
nearly neutralized with HC1, and were at temperatures of 22 to 24° C.
72
-------
The highest concentration causing no appreciable reduction of the
heart rate was not determined, but could well have been as low as 10~5 M
(0.26 mg/1 as CN).
Armstrong and Fisher (1940) found that 68 percent of the normal heart
beat frequency of embryos and young larvae (sac fry) of the Atlantic
salmon, Salmo salar, was cyanide-stable. The embryos used in the tests
were apparently all approaching the hatching stage, for the authors
stated that "if hatching had not yet occurred naturally, the egg
membranes were removed to facilitate examination of the heart." The
observations were made at a temperature of 6° C and pH 7.1. The highest
cyanide concentration having no effect on the cardiac rhythm again was
not determined, but clearly was far below 10~5 M. The authors stated
that "the greatest reversible inhibition of the frequency which can be
produced by cyanide does not result in complete stoppage of the heart."
This statement is not highly instructive and its meaning is not very
clear. It seems to imply that an irreversible reduction of the heart
rate by more than 32 percent (the portion reported to have been cyanide-
sensitive) and leading to complete stoppage of the heart could be pro-
duced by exposure of the test animals to cyanide, but the lowest cyanide
concentrations causing the maximum reversible (32%) inhibition and the
greater, irreversible inhibition and stoppage of the heart beat were not
reported.
Brinley's (1930) data pertaining to mummichog, Fundulus_ heteroclitus,
embryos indicate that at a very high concentration of KCN (1/10 M),
stoppage of the heart, which occurred after exposure for about 1 hour,
was not preceded by any marked reduction of the rate of heart beat.
The rate decreased to decidedly lower levels at somewhat lower concen-
trations of KCN, at which the heart continued to beat for much longer
periods (4-7 hours). The pH of Brinley's test solutions, which were
73
-------
prepared with, slightly alkaline sea water (pH about 8.2) was not
adjusted, and the most concentrated KCN solutions were said to have
been strongly alkaline; their unduly high pH may well have influenced
the test results. It is evident, however, that the time to stoppage
of the heart can be independent of, or can even vary inversely with,
the heart beat frequency at the time immediately preceding the stoppage.
The cessation of beating of the heart observed by Brinley at high
cyanide concentrations was not irreversible. Within about half an hour
to 4 hours after return to clean sea water of exposed embryos whose
hearts had ceased beating for as long a period as 60 to 180 minutes,
the heart beat was resumed, the time required for this recovery
increasing with increase of the cyanide concentration to which the
embryos had been previously exposed.
TESTS FOR OTHER NONLETHAL INJURY
In experiments on the influence of free cyanide (NaCN) on the behavior
of cichlids, Cichlasoma bimaculatum, in laboratory aquaria, Brockway
(1963) observed no effect on the reproductive activity of adults nor on
the schooling and fright reactions of juvenile fish at concentrations
of 0.02 and 0.10 mg/1 as CN. At the higher one of these two test con-
centrations, which is not far below the 48-hour median tolerance limit
for the juveniles, determined to be about 0.135 mg/1 as CN at the
experimental temperature of 25° C, the young fish were, however, found
to feed with less vigor and to consume food slower than did the controls.
At the lower concentration tested, this effect was not seen.
Brockway performed some preliminary experiments on effects of chronic
cyanide (NaCN) poisoning also on enzyme activities in liver tissues of
the cichlids. His results indicated reduction of cytochrome oxidase
and succinic dehydrogenase activities by about 10 and 70 percent,
respectively, and increase of aldolase activity by some 263 percent in
74
-------
liver tissues of cichlids that had been exposed for 60 days to a con-
centration of 0.10 mg/1 as CN. However, a subsequent, more thorough
investigation by Leduc (1966) revealed no simple relations worthy of
note between cytochrome oxidase, aldolase, or peroxidase activities in
liver tissue homogenates and the cyanide concentrations (nil and 0.02
to 0.1 mg/1 as HCN) to which the cichlids had been exposed for periods
of 1 to 36 days. Some large variations of the enzyme'activities that
were observed are not easily interpreted because of lack of consistency.
Succinic dehydrogenase activities were not evaluated. Considerable
increases of the proteolyttc activity of homogenates of intestinal
tissues of cichlids exposed to cyanide were observed. The proteolytic
activity values tended to increase fairly consistently with increases
of both the exposure concentrations (to 0.10 mg/1 as HCN) and the
duration of exposure (to 20 days). There may well be some connection
between these increases of digestive enzyme activity and the relatively
high food consumption rates of cyanide-exposed cichlids that have been
already reported.
Leduc (1966) also determined changes in body weight and composition of
unfed cichlids exposed to cyanide (NaCN) levels of 0.02, 0.04, and
0.09 mg/1 as HCN and of unfed controls. Comparison of the changes
observed after 6, 12, and 24 days of starvation revealed a marked
acceleration of the loss of energy reserves (computed caloric values)
in the presence of cyanide. This effect was most pronounced at the
highest cyanide concentration tested and after only 6 days of starva-
tion, when the total or cumulative loss (in kilocalories per gram of
mean, dry body weight) at the highest concentration was found to have
been about four times the control value. After 12 days of starvation
at that concentration, however, the cumulative loss was computed to have
been less than twice the control value, and the difference in cumulative
loss of energy reserves of the experimental and control fish had
virtually disappeared after 24 days of starvation. The differences
75
-------
observed early in the experiment strongly indicate loss of metabolic
efficiency of the cyanide-poisoned fish, as did also Leduc's data on
food conversion already reported. The significance of the subsequent
disappearance of the differences is not entirely clear, but it can be
attributed to adaptation and probable reduction of activity of the
poisoned and starved fish, whose metabolic rates evidently declined
markedly during the last 12 days of the experiment.
In tests of short duration (100-155 minutes), Jones (1947) noted pro-
nounced reductions of the rate of oxygen consumption of threespine
sticklebacks exposed to NaCN solutions at 17° C and pH 7.0 but not yet
overcome by the poison. In 3-4 x 10 M solutions (i.e., at concen-
trations of 0.78.-1.04 mg/1 as CN), average oxygen consumption rates
lower than normal rates by 45-68 percent were reported to have been
observed after 90-155 minutes of exposure. The 4 x 10" M concentra-
tion, tolerated for 100 minutes before conclusion of the test, was
said to be a "critical" one, but even the 3 x 10 M concentration,
tolerated indefinitely by few if any species of fish, actually is far
above the minimum lethal level for the threespine stickleback according
to the data of Costa (1965). Opercular or breathing rates of Jones'
sticklebacks exposed to the NaCN solutions increased at first but later
fell to values below the normal values as the oxygen consumption rates
continued to decline. Increases of the opercular rate of fish exposed
to cyanide solutions have been noted also by Ishida (1947). Remarkably
rapid recovery upon return to clean water of fish that had almost ceased
breathing in rapidly lethal cyanide solutions was noted by Jones, as it
had been also by Karsten (1934) in his experiments with trout.
Carter (1962) reported that, in 300-minute tests, the oxygen consump-
tion of fingerling brown trout that had been confined in sealed bottles
until they died was found to have been measurably reduced at cyanide
concentrations as low as 0.025 mg/1. Negilski 0-973), using a fairly
76
-------
refined technique of oxygen uptake measurement (continuous-flow res-
pirometers), found that the oxygen consumption rate of juvenile chinook
salmon, Oncorhynchus tshawytscha, was depressed by about 31 percent, on
the average, in the presence of 0.02 mg/1 of free cyanide a's CN, or
about one-fifth of the 96-hour median tolerance limit for these fish.
However, the measurements were few, their results highly variable, and
the differences between the mean values for the experimental and control
fish not highly significant statistically (probability level 0.83).
Brockway (1963), using essentially the same method, was unable to
demonstrate a definite effect of free cyanide concentrations ranging up
to 0.08 mg/1 as CN on the also highly variable oxygen consumption rates
of his juvenile cichlids, Cichlasoma bimaculatum.
Negilski's (1973) observations on the direct or indirect influence of
free cyanide on the food supply and the production of juvenile chinook
salmon in artificial streams with model animal and plant communities
set up in the laboratory are not very instructive. The desired, con-
stant cyanide concentrations evidently could not be maintained in the
circulated water, and the average levels to which the organisms were
exposed are not known. Salmon production apparently was increased or
reduced somewhat in different tests by the addition of small amounts of
cyanide to the water, but, doubtless because of losses of the toxicant,
only minute amounts of free cyanide were detected in the stream water
by the few chemical analyses that were done. Because the significance
of the experimental results is too obscure, they have not been con-
sidered in connection with the discussion of the effects of free cyanide
on the growth of fish and nothing more will be said about them here.
Chan (1971) found that previous exposure of yearling rainbow trout for
28 days to five free cyanide concentrations ranging from 0.01 to 0.037
mg/1 as HCN at 10° C affected their subsequent osmoregulation. There
were no noticeable effects on the early phases of adaptation of the fish
77
-------
to a water salinity of 18.9 parts per thousand, but after exposure to
the saline medium for 260 hours, the cyanide-poisoned fish were found
to have plasma osmotic pressures and chloride concentrations higher
than those of controls. When some remaining fish were transferred from
the saline medium back to fresh water, again there was no obvious, early
effect, but by the end of the test all the cyanide poisoned fish had
plasma osmotic pressures and chloride concentrations lower than those
of controls. Similar results were obtained when trout adapted for 376
hours to the saline water were transferred to fresh water containing
free cyanide at concentrations ranging from 0.01 to 0.037 mg/1 as HCN.
Histological examination revealed some effect of exposure of the fish
for 28 days to 0.037 mg/1 of free cyanide as HCN on the epithelial cells
of the thyroid gland (5.3 percent reduction of cell height). Exposure
of the saline water induced a 60.8 percent increase of the height of
epithelial cells of the thyroid follicles in controls, whereas the
cyanide-poisoned fish responded to the same treatment by only an 8.6
percent increase of the epithelial cell height.
Hiatt, Naughton, and Matthews (1953b), who observed effects of various
chemicals on the behavior of the marine fish Kuhlia sandvicensis during
the first 2 minutes of exposure, reported the irritant activity of KCN
to have been violent, moderate, and slight at levels of 10, 1.0, and 0.1
mg/1, respectively (4-0.04 mg/1 as CM). Elsewhere (Hiatt, Naughton, and
Matthews, 1953a), these authors reported only a slight reaction of the
same fish to NaCN at a concentration of 1.0 mg/1 (0.53 mg/1 as CN), but
a violent reaction at a concentration of 2.0 mg/1 was reported also.
AVOIDANCE REACTIONS
Avoidance reactions of fish to cyanide in a horizontal glass tube, one
half of which contained flowing (continuously renewed) tap water and the
other half a flowing solution of NaCN (5 x 10~3 N to 1 x 10~6 N) have
78
-------
been studied by Costa (1965). He reported some demonstrable but not
sharp avoidance of a concentration as low as 1(T6 N, or 0,026 mg/1 as
CN, by young brown trout, wjiich proved most responsive, and also by the
minnow Phoxinus phoxinus and the threespine stickleback, but not by
young eels (elvers), Anguilla anguilla, nor by young goldfish., Carassius
auratus. The sticklebacks showed only a vague and very long delayed
reaction ;to this concentration. Avoidance of 10 N solutions, or 0.26
mg/1 CN, was reported to have been shown by all of the above species,
but that shown by the elvers was very slight. Definite avoidance by
c
all the species of 5 x 10 N solutions, or 1.3 mg/1 CN, was observed.
However, the initial avoidance reactions even to much higher concentra-
tions were not immediate, and avoidance of low but still eventually
fatal concentrations usually was long delayed. After withdrawing from
a dilute cyanide solution, fish often reentered it and remained in it
for some time before withdrawing again.
In Costa1s experiments, young individuals generally reacted more slowly
than did older and larger fish, but they displayed a greater ability to
withdraw from cyanide solutions before being overcome by the poison;
only 5 x 10 N NaCN solutions were used in the comparative tests with
groups of fish of different size. In the presence of a fairly high
concentration of sodium thiosulfate (0-025 N) , the avoidance by fish of
low concentrations of free cyanide was less pronounced or slower than
it was in the absence of the antidote. Prolonged prior exposure of fish
to thiosulfate also rendered them less responsive to the low cyanide
levels. The reactions of threespine sticklebacks to cyanide were more
rapid at high temperatures than at low temperatures. Reduction of the
pH of a 5 x 10~5 N NaCN solution to pH 5.5 or less by addition of HC1
reduced the reaction time and the amount of time spent by the stickle-
backs in the solution, as did also reduction of the dissolved oxygen
content of the solution. Progressive increase of the pH of the solution
to different values up to 10.1 by addition of NaOH progressively delayed
79
-------
and reduced or completely eliminated the avoidance reactions; however,
at pH 11.3, the reactions were more rapid and pronounced than they were
at pH 10.1 and much like those observed when the pH of the solution was
9.1. Since the pH of the cyanide-free water was not altered in these
experiments as was that of the cyanide solution, it was not possible to
distinguish between reactions to cyanide and reactions to the pH differ-
ences themselves. Avoidance of the solutions of very low and very high
pH may well have been mostly avoidance of the extreme pH values or of
high carbon dioxide levels and not of cyanide.
Summerfelt and Lewis (1967) performed experiments with a different kind
of apparatus, a trough 6.4 m long and 0.6 m wide divided into six com-
partments of equal length with five transverse gates that could be
simultaneously opened or shut (raised or dropped). With the gates shut,
five green sunfish, Lepomis cyanellus, averaging 10.8 cm in length were
placed in each compartment and the toxicant was introduced into the
compartment at one end of the trough. After 20 minutes the gates were
opened, and after 5 or 10 (or 15; statements in the text of the paper
are contradictory) additional minutes they were again closed, trapping
the fish in the compartments in which they were at that time. The
final distribution of the fish in the trough, or among the compartments,
which was said to have been uniform in control tests without any chemical
added to the water in the trough, then was determined and the degree of
avoidance of the toxicant by the fish thus was evaluated. The tempera-
ture of the water in the trough averaged about 23° C, the pH averaged
7.2, and the dissolved oxygen concentration was rather low, averaging
4.5 mg/1. Summerfelt and Lewis reported that NaCN "was found to be
moderately effective as a repellent at 5 mg/1 and to produce an avoid-
ance response at 1.0 mg/1." No response was observed to a concentration
of 0.5 mg/1 or less. Inasmuch as green sunfish have been reported
killed in less than 3 hours at a concentration of 1.0 mg/1 (as NaCN) and
temperatures of about 12 to 27° C, and in 4 to 6 hours at the concentra-
tion of 0.5 mg/1 and temperatures of 26 to 28° C (Bridges, 1958), it
80
-------
appears that decidedly lethal levels did not all prove even "moderately"
repellent. It should be noted, however, that the reported concentrations
were the initial concentrations in the compartment at one end of the
trough into which the toxicant was introduced. Since there must have
been some mixing and dilution of the solution in this compartment with
water from the adjacent compartment, producing a concentration gradient
in the trough, the exact concentrations to which the fish responded or
that they failed to avoid are not known.
Summerfelt and Lewis also briefly reported a pertinent observation made
in the field in connection with the sampling of a fish population by the
use of cyanide to poison the fish. They stated that a "0.3-acre (0.12-
ha) cove of a lake was screened off with a seine and 1.0 mg/1 cyanide
was applied opposite the screen." They further stated that the fish of
all species present in the area "exhibited a strong tendency to congre-
gate along the screen in proximity to the untreated water." It is not
entirely clear just what was meant by the first statement quoted.
Probably the chemical NaCN was somehow more or less uniformly distri-
uted throughout the screened-off cove in amount sufficient to produce a
concentration of NaCN (not of cyanide as CN) of 1.0 mg/1 in all of the
water then in that cove. However, the puzzling statement in question
can be variously understood, and one cannot say what free cyanide con-
centrations the responding fish actually had encountered or been exposed
to in the treated water.
From the last-mentioned observation and the results of Costa1s experi-
ments, it seems reasonable to conclude that fish can sense the presence
of harmful concentrations of free cyanide in their medium before they
are overcome by the poison, and they probably are able usually to escape
from lethal concentrations present only in the immediate vicinity of
waste diffusers or other outfalls (i.e., where the poison is diluted to
tolerable levels within a very short distance from the points of
81
-------
discharge). However, cyanide concentrations can increase very gradually
with, decreasing distance from a source of heavy contamination of water
with the poison. In view of the slow and uncertain reactions observed
by Costa and by Summerfelt and Lewis in the laboratory, one should not
conclude that fish encountering harmful cyanide concentrations in such
a gradient will usually avoid fatal exposure to these concentrations by
turning and swimming away. When confined in a tube such as that used by
Costa, with a rather sharp boundary or narrow transition zone in the
middle between contaminated and uncontaminated waters, a fish can
respond or behave in one of three ways only. It can remain all or most
of the time in a solution with a given, uniform concentration of a
toxicant, it can remain all or most of the time in uncontaminated water,
or it can move back and forth indiscriminately, spending about the same
amount of time in each half of the tube. If it is active, it must turn
and reverse the direction of swimming frequently, and it cannot choose
to continue swimming in one direction and into progressively increasing
concentrations of the poison; a flight reaction without frequent turning
can lead only to immediate escape from the unfavorable medium. After a
number of excursions into, and sojourns in, the contaminated water, or
even a single, highly distressing exposure to such water, the fish may
learn to avoid that portion of the tube where it has recently suffered
distress. Such avoidance of cyanide in the tube, or similar behavior
i .'
of fish confined in a trough, does not signify that the fish would not
continue swimming away from a zone of clean water or in the direction
of gradual increase in concentration of the toxicant until it is over-
come by the poison, if it had the opportunity to do so. Effectiveness
of cyanide as a "directive stimulus", or its ability to repel fish
from large areas of natural habitat polluted with eventually lethal or
sublethal amounts of the poison, has not yet been convincingly demon-
strated. Ishio's (1965) sketchily reported results of experiments with
a gradient tank (concentration gradients), indicating only a 50 percent
"frequency of avoidance" by fish of a lethal (10~5 M) level of HCN, do
not in any way contradict this conclusion (see Doudoroff, 1965).
82
-------
SECTION V
TOXICITY OF COMPLEX CYANIDES
GENERAL, CHEMICAL BACKGROUND
Simple cyanides of some heavy metals, such as mercuric cyanide, Hg(CN)2,
as well as those of alkali and alkaline earth metals, are highly soluble
in water. Those of most of the heavy metals likely,to be important
components of cyanide-bearing wastewaters, however, are only,very
slightly soluble or almost insoluble, cadmium cyanide,, Cd(CN)2, being
an important exception. These almost insoluble ecyaaiiMss,, such as those
of silver or nickel, AgCN.or Ni(CN)2, may be farmea.*isifeenssoluble salts
of the metals, such as silver nitrate, AgNO^, or..Trirfekdl ssulfate, NiSO,,
are combined in solutions with simple, alkali metal cyanides, such as
KCN or NaCN. However, they are formed and remain :as permanent precipi-
tates only when the amounts of the heavy metals exceed the amounts
capable of reacting with the alkali metal cyanides to rform certain
soluble, complex metallocyanides (double salts), such as potassium silver
cyanide, KAg(CN)2j or sodium nickelocyanide, Na2NiCCN)4. The ionization
or dissociation of these highly soluble complex cyanides yields alkali
metal cations and complex metallocyanide anions, such as Ag(CN)2~ or
Ni(CN)4~2, which, in turn, dissociate. The dissociation of the complex
ions, unlike the first dissociation mentioned, may be very slight or
incomplete; it yields metallic cations and free cyanide ions, and the
hydrolytic reaction with water of the cyanide ion deriving from the dis-
sociation yields molecular HCN. The following two equilibrium equations
83
-------
pertain to the dissociation of the nickelocyanide or tetracyanonickel-
ate(II) complex:
= KD = about 5 x 10~31
and — - = Ka = about 5 x 1(T10
CH+] [ON']
[HCN]
where Kp is the cumulative dissociation or instability constant for the
?
tetracyanonickelate(II) complex ion, Ni(CN)4~ , and Ka is again the
ionization constant for HCN. The value given above for each of the
constants is an approximation, or rounded average, of reported values
deemed most reliable and pertaining to temperatures of 20-25° C. The
bracketed symbols again represent molar concentrations or, strictly
speaking, activities of solution components. It should be understood
that the two equilibria considered above "compete" for the cyanide ion,
which appears in both equations, and that both equations must be
satisfied when the entire system is at equilibrium.
As the nickelocyanide complex ion dissociates, a preponderant portion
of the cyanide ion (CN~> that is liberated is converted into molecular
HCN at the pH values of most natural waters. This removal of the CN~
ion from the system permits the dissociation of the complex to proceed
much farther than it could otherwise, until no more HCN can be formed.
At the ordinary pH levels, the amount of nickel ion (Ni ) liberated,
in gram atomic weights per liter, is almost equal to, or only slightly
greater than, one fourth of the amount of HCN formed, in gram molecular
weights (moles) per liter, and thus will be nearly constant as long as
the HCN concentration remains unchanged. A reduction of the hydrogen
ion concentration by one half (i.e., an increase of pH by only slightly
more than three tenths of a pH unit) requires a doubling of the CN" ion
84
-------
concentration if the molecular HCN level is to remain unchanged,
according to the second equation given above. Two raised to the,, fourth
power (see the first equation above) is 16. Therefore, if the molecular
HCN level is to remain constant in solutions of the nickelocyanide com-
plex in which the total cyanide present is very preponderantly in the
form of the complex (i.e, , in, slightly alkaline and not exceedingly
dilute solutions) , the . concentration of Ni(CN)4~2 ion must increase
about 16-fold C not exactly .that much) when the pH value is increased
by 0.3. For the silver- cyanide complex ion, Ag(CN)2~, the corresponding
increase is about a fourfold increase (x 2^).
»,
In nickelocyanide solutions, the NiCCN)3~ ion species is not formed in
appreciable amounts. When the CN/Ni mole ratio is less than 4.0 Ce-g->
when nickel ion, has been added to a solution of Na2Ni(CN)4 containing no
additional cyanide), the very slightly soluble cyanide of nickel, Ni(CN)2
•Z J
is formed instead. The complex ion species Ni(CN)g , AgCCN)^" , and
are known but are unstable in dilute, solutions, and they can
be formed and can persist indefinitely only in the presence of, much free
2
cyanide. On the other hand, the cuprocyanide ions Cu(CN)2~ and CuCCN)^
can continue to coexist in dilute solutions, at equilibrium in consider-
able amounts and in the presence of relatively very little or much free
cyanide, their relative amounts depending on the CN/Cu mole ratio and
other factors. _. ,
Referring only to the two equations considered above, one can readily
calculate the approximate concentration of molecular HCN that is to be
expected at equilibrium in a dilute solution of Na2Ni(CN)4 with any
known total cyanide content and pH, and at a temperature of 20-25° C.
The total cyanide level in such a solution that is required for producing
at equilibrium any desired level of molecular HCN at a given pH can be
estimated likewise. The computed values may not prove entirely correct
because of :some minor complications or sources of error that are not all
85
-------
fully known or understood, such as an influence on the equilibria of the
ionic strength of the solutions and some possible formation of metallic
ion species (hydrolyzed nickel species) other than the simple, solvated
metal cations (aquonickel). "Apparent KD" values computed by Broderius
(1973) from determined molecular HCN levels in various nickelocyanide
solutions of different total cyanide content and pH may be more reliable,
or more useful when calculations pertaining to similar solutions are
made, than other reported Kp values that thermodynamically may be
strictly or more nearly correct.
Calculations pertaining to the dissociation of the monovalent dicyano-
argenate(I) complex ion, Ag(CN>2~, in waters of low chloride content
are not obviously more complicated than those considered above. However,
for reasons not yet determined, there is somewhat more serious disagree-
ment between the "apparent KQ" values for the ion computed by Broderius
(1973), and comparable values otherwise derived (and perhaps inaccurate)
that had been previously reported in recent chemical literature. Much
greater discrepancies between corresponding values for both the ferro-
eyaaside ion, Fe(CN)6 , and the ferricyanide ion, Fe(CN)6~3, also are
still unexplained, but for reasons to be considered later, this is a
matter of academic interest only. It should be understood that
Broderius1 calculations of apparent KJJ values are based upon the
assumption of the simplest possible systems, like that to which the
foregoing equilibrium equations pertaining to the dissociation of the
nickelocyanide complex apply. As indicated already, however, the
systems in questionsactually may be much more complicated. For example,
protonated species; oftthe cyanide complexes, such as HEe(CN)g~ or
-2
H2Fe(CN)6 , may occur in significant amounts. There is not now suffi-
cient reason to conclude, therefore, that the published, thermodynamie
constants (Kg values) for the ferrocyanide and ferricyanide complex ions
are grossly erroneous; there are reasons to believe otherwise and to
seek in the complexity of the systems involved a different explanation
86
-------
for the striking lack of agreement of these values and Broderius1 appar-
ent or formal constants. A tendency of silver ions to combine with
chloride ions (which thus compete with cyanide ions for the metal
cations) in solutions of high chlorinity introduces another interesting
complication to be more fully considered later, together with informa-
tion on the toxicity of the silver-cyanide complex.
Because of the indicated nature of the stepwise dissociation of cupro-
cyanide complex ions, calculations of equilibrium levels of the various
ion species and molecular HCN in cuproeyanide solutions are obviously
much more involved and difficult than those pertaining to the niekelo-
cyanide complex. Constants pertaining to the dissociation of the dif-
ferent, coexisting, complex (cuprocyanide) ion species must be intro-
duced into the calculations, which cannot be adequately dealt with here.
When the CN/Cu mole ratio is 4.0, for example, use in the calculations
of the KQ value for the Cu(GN)4 ion and of only two equations corres-
ponding to those given for equilibria in nickelocyanide solutions with
the same mole ratio of cyanide to metal leads to patently wrong answers.
Some authors apparently have overlooked the error involved in such
improper calculations and have reached incorrect conclusions partly for
this reason. Calculations pertaining to systems with CN/Cu mole ratios
of 3.0 or 2.0 (Broderius, 1973) also are not simple. It appears that
the very stable Cu(CN)2~ ion can be predominant in dilute solutions with
CN/Cu mole ratios well in excess of 2.5 when equilibrium has been
attained, and this ion can persist at ratios above 3.0. But even a
o
CN/Cu mole ratio of 2.0 does not ensure that virtually no CuCCNJj"'6 ion
and no undissolved cuprous cyanide (CuCN) will be present at equilibrium
when CuCN is dissolved in a dilute solution of NaCN.
The reader should understand that the same value for the cumulative
dissociation or instability constants of different complex ions does not
signify equal stability of the complexes if the mole ratios of cyanide to
87
-------
metal are different. Thus, a KQ of 10~21, or even 10~19, for an ion
such as the AgCCN)2~ ion signifies high Stability of the complex, but
a KD value of-10"22 for the NiCCN)^2 ion (an erroneous value long
accepted by chemists), although smaller, would signify a relatively low
level of stability. This is true because in the first equilibrium
equation given above, p'ertaining to the nickelocyanide complex, the
value of [CN~] is raised to the fourth power, whereas in the corres-
ponding equation pertaining to the silver-cyanide complex it must be
only squared.
The time periods required for attainment of equilibria upon dilution of
solutions of the more stable metallocyanide complexes, and also when
different metal salts and free cyanide are combined to form the complexes,
are extremely variable. Broderius (1973) determined the concentrations
of molecular HCN in various buffered experimental solutions of varying
pH and total cyanide and heavy metal content and studied the' rates of
their change attributable to dissociation or formation of metallocyanide
complexes. He found that both the dissociation and the formation of the
nickelocyanide complex were rather slow in nearly neutral and acid
solutions with low or moderate total cyanide concentrations (50 mg/1 or
less). Dissociation of the complex already formed by combining concen-
\
trated solutions of NaCN and NISC^ was especially slow. When the
relatively concentrated, highly alkaline solutions of the complex were
greatlyrdiluted with buffered water, with consequent reduction of pH,
more time was usually required for attainment of equilibrium than was
required when the salts were added separately to the diluent to form
the complex (i.e., were combined in dilute solutions). The difference
was most pronounced when the pH and the cyanide concentrations were low.
At pH 6.5 and total cyanide concentrations of 0.5 and 5.0 mg/1, dissocia-
tion equilibria were attained only after about one week to 10 days. The
time required for attainment of equilibrium decreased with increase of
pH or of the total cyanide concentration, being apparently directly and
-------
linearly related to the percentage of the total cyanide that is present
as HCN at equilibrium.
After enough nickel to combine with all of the cyanide present, forming
the Ni(CN)4~2 complex, had been added (as NiS04) to a buffered NaCN
solution with a cyanide content as high as 250 to 500 mg/1 as CN, the
concentration of molecular HCN was found to be curiously low initially
and to increase with time until equilibrium was attained. In similar
but more dilute mixtures of NaCN and NiS04, the HCN concentrations
always were relatively high initially and decreased with time, as they
had been expected to decrease because of progressive formation of the
_2
Ni(CN)4 complex. The anomalous increase of the HCN concentration in
the more concentrated solutions of NaCN combined with NiS04 was tenta-
tively explained by Broderius as having been probably due to gradual
release.of cyanide from small amounts of complexes such as the
"7
pentacyanoniekelate (II) complex ion, Ni(CN>5 . These complexes were
supposed to have been rapidly formed initially, together with the much
more abundant and stable Ni(CN)4 ion, under the conditions of the
experiments in question.
The formation and the dissociation of the silver-cyanide complex,
Ag(CN)2~, were found to be quite rapid, the HCN levels determined within
a few hundred minutes after preparation of all tested solutions having
closely approximated the equilibrium levels. On the other hand, disso-
ciation of the iron-cyanide complexes, Fe(CN}6~4 and Fe(CN}6~3, proved
exceedingly slow, storage of some solutions of potassium ferro- and
ferricyanide for 140 days or more in the dark having been required far
attainment of equilibria. The rate of dissociation of the euprocyanide
complex ion Cu(CN)2~, like that of the nickelocyanide complex, was found
to vary greatly with pH and total cyanide concentration. At a total
cyanide level of 5 mg/1 and pH 6.5, equilibrium was attained only after
about 3 weeks, whereas at a total cyanide concentration of 50 mg/1 and
pH 7.5, equilibrium was attained in less than 1 day.
89
-------
The appearance of a precipitate (turbidity) upon the addition of a
solution of a metal salt to a cyanide solution does not signify that no
free cyanide at all remains in the mixture. Some free cyanide deriving
from the dissociation of a cyanide complex can coexist with a precipi-
tated, solid metal cyanide. A heavy metal that forms a complex of
relatively low stability can be driven out of the complex by a metal
capable of forming a more stable cyanide complex. Thus, when nickel
_2
ion is added to a solution containing the zinc-cyanide complex, Zn(CN)4
•I rj
(Kp about 10 ), the zinc may be replaced by nickel in the complex, and
the consequent liberation of zinc ion may cause precipitation of zinc
cyanide.
;
The iron-cyanide complexes are very stable in the dark, but are subject
to rapid photodecomposition with liberation of the cyanide. Reactions
of free cyanide with cupric ion are complicated, involving reduction of
copper(II) to copper(I) as well as other intricacies. As noted already,
competition for metal ions of cyanide ions and other anions that can
combine with the metals can influence the dissociation of metal-cyanide
complexes, promoting the liberation of cyanide from the complexes. The
chemistry of the complex cyanides obviously is not simple, and it is
possible to provide here only a very sketchy and not thoroughly instruc-
tive introduction to the subject.
TOXICITY OF THE METALLOCYANIDE COMPLEXES IN GENERAL
Early studies of the toxicity to fish of the metal-cyanide complexes
have shown clearly that cyanide combined with heavy metals to form at
least some of the more stable complexes is nontoxic or much less toxic
than free cyanide. Potassium ferrocyanide, K4Fe(CN)6, was for a long
time considered to be a relatively harmless water pollutant (Ellis,
1937). Later, the possibility of production of enough free cyanide to
kill fish by photodecomposition of very small amounts (less than 2 mg/1)
90
-------
of the iron-cyanide complexes was recognized (Burdick and Lipschuetz,
1950). Milne (1950) noted that a high concentration of the nickelo-
cyanide complex (more than 100 mg/1 as CN) was not evidently toxic to
goldfish in a moderately alkaline medium. He then apparently assumed
that complexation of cyanide with any heavy metal would result in
effective detoxification of the cyanide, rendering it harmless in any
receiving water. Evidently he failed to consider adequately, however,
the large differences in stability of the different complexes and the
important influence of pH on the dissociation of the complexes in dilute
solutions.
Stumm, Woker, and Fischer (1954) observed much reduction of the toxicity
of cyanide to minnows, Phoxinus laevis, when it was complexed with
nickel, copper, or iron, but not when it was combined with zinc or
cadmium. They realized that the very pronounced toxicity of the zinc-
cyanide and cadmium-cyanide complexes is attributable to their rela-
tively low stability, and that free cyanide or HCN could have been the
most important lethal agent in tested solutions of these and other
cyanide complexes. An attempt was made, with some apparent success, to
relate effective exposure times to computed HCN concentrations in the
various solutions of complex cyanides. However, as Doudoroff, Leduc,
and Schneider (1966) later noted, the ionization constant for HCN was
not correctly introduced into the computations of free HCN levels. This
error compensated for the use in the computations of a value for the
—??
dissociation constant of the nickelocyanide complex (10~ ) that later
proved grossly inaccurate, but because of several sources of error, none
of the computed HCN levels were correct. Although a general, inverse
relation between the toxicity of the metal-cyanide complexes and their
stability was, to some extent, revealed, direct dependence of the
toxicity of solutions of the complexes on their free cyanide or HCN
content was neither demonstrated nor disproved. At least some of the
toxicity of solutions of the more stable complexes that were not very
91
-------
rapidly fatal was vaguely attributed to the presence of free heavy
metal ions, but no convincing evidence of the suggested, important
role of free metallic cations was presented.
Doudoroff (1956), experimenting with fathead minnows, Pimephales
promelas, also found the toxicity of the zinc-cyanide and cadmium-cyanide
complexes to be very great. He concluded that this high toxicity was
ascribable mostly to liberation of virtually all of the cyanide (i.e.,
nearly total dissociation of the complexes) in very dilute solutions.
Complexation of cyanide with nickel, copper, and iron greatly reduced
its toxicity, but the toxicity of the nickelocyanide complex was found
to vary strikingly with the pH of the solutions and to be quite pro-
nounced (not very much less than that of NaCN) at pH levels as low as
6.5-6.6. The detailed observations on the influence of pH on the
toxicity of solutions of the nickelocyanide complex were shown generally
to support the hypothesis that this toxicity is ascribable to, or is
dependent chiefly on the concentration of, molecular HCN. However, some
features of the experimental results could not be adequately''explained
on the basis of the available chemical information and thus fully
reconciled with the above-stated hypothesis. Serious inaccuracy of the
then generally accepted cumulative dissociation constant for the
— 22
nickelocyanide complex (10 ), later demonstrated also by chemical
studies of other investigators, was the probable reason advanced by
Doudoroff for some of the disagreement between his observations and
expectations based on theory. A nearly correct estimate of the constant
(about 10" ), derived by computation from the toxicity data, then was
offered.
Bucksteeg and Thiele (1957) compared the toxicities to fish (species
unknown) of cyanide complexes of zinc, cadmium, copper, and nickel with
the toxicity of KCN. Concentration limits for harmful action on fish of
KCN, K2Zn(CN)4, K2Cd(CN)4, K3Cu(CN)4, and K2Ni(CN)4 were said to be 0.1,
92
-------
0.3, 0.75, 1.0, and 30 rag/1 as CN, respectively. Considerable reduction
of the toxicity of cyanide upon its complexation with all of the four
metals is indicated. The publication is of little value, however,
because experimental procedures and conditions, such as test solution
pH values, are not reported there. The reported concentration limits
could not be very meaningful or reliable even if they were correct
values for some particular level of pH, inasmuch as the degree of dis-
sociation and toxicity of a metallocyanide complex can vary greatly with
the pH of the water. But some of these limits, such as the inordinately
high value of 0.75 mg/1 as CN for the cadmium-cyanide complex, also are
not at all in agreement with other published toxicity data and with
pertinent theoretical considerations.
Using the analytical method of Schneider and Freund (1962), Doudoroff,
Leduc, and Schneider (1966) were the first investigators to relate the
toxicity of solutions of complex cyanides, as well as of NaCN, to
analytically determined levels of molecular HCN in these solutions. The
tested solutions of NaCN and of a number of different metal-cyanide
complexes, of which the nickelocyanide complex was tested most often,
varied widely in both total cyanide content and pH. They were prepared
with a stream water. The recorded median resistance (immobilization)
periods for bluegills at 20° C were plotted in a graph against the
determined molecular HCN levels, and also against the concentrations of
i
total cyanide added in preparing the solutions (expressed as HCN).
Generally good correlation was found between the resistance time and the
molecular HCN concentration, all but one of the plotted points pertaining
to this relationship falling close to a straight line which was fitted
to these data. The single exception was a point representing the result
of a test of a slightly alkaline solution of the silver-cyanide complex.
This discrepant result strongly indicated considerable toxicity of the
silver-cyanide complex ion itself, the fish having died almost as soon
at an HCN concentration that clearly could not alone have been fatal as
93
-------
they died at a much higher HCN level but the same concentration of the
complex (1° mg/l &s CN) and a lower pH. Pronounced toxicity of the
cuprocyanide complex ion Cu(CN)2~> now believed to be even more toxic
than the Ag(CN)2~ ion (Broderius, 1973), was not revealed. The tested
solutions containing copper were prepared by combining NaCN with cupric
sulfate, CuSO^, which react to produce cuprocyanide ions. The mole
ratio of cyanide to copper (CN/Cu) is believed to have been well above
2.0 (between 2.0 and 2.5), and much free cyanide was present, masking or
rendering unimportant any effects of the less toxic Cu(CN)2~ ion, even
though most of the cyanide present must have been in the form of this
complex.
No correlation was seen between median resistance time and total cyanide
concentration, except that all the points representing results of tests
of NaCN solutions and solutions of the zinc-cyanide and cadmium-cyanide
complexes were distributed along the above-mentioned straight line.
Proximity of these points to the line is not meaningful, inasmuch as the
total cyanide concentrations in the particular solutions in question
differed only slightly from the concentrations of molecular HCN.
The findings of Doudoroff, Leduc, and Schneider (1966) did not show thev
toxicity of the metallocyanide complex ions themselves to be always
slight, or the contribution of these ions and of the more toxic heavy
metal cations deriving from their dissociation to the toxicity of solu-
tions of complex cyanides to be always negligible. Indeed, pronounced
toxicity of the silver-cyanide complex ion was revealed. Nevertheless,
these experimental results did lend strong support to the supposition
that the acute toxicity to fish of very dilute solutions of complexes
often to be found in industrial waste waters is usually due predominantly
to the presence of free cyanide liberated by dissociation or decomposi-
tion of the complexes in the form of accurately measurable, molecular
HCN. Therefore, it now seems reasonable to conclude that the toxicity
94
-------
of waters polluted only with alkali metal cyanides or metallocyanides
(complex cyanides), and not additionally with, metallic or other toxi-
cants in considerable amounts, can usually be estimated without very
serious error by measuring and considering only their molecular HCN
content. In this connection, it can be noted that silver is not likely
to be wasted in large amounts, and that the CN/Cu mole ratios in metal-
finishing (plating) wastes containing cuprocyanide complex ions, except
effluents effectively treated for cyanide removal, are usually well
above 2.0, as they were also in solutions tested by Doudoroff, Leduc,
and Schneider.
Views contrary to the conclusion stated above have been expressed,
notably in a brief and incomplete review of pertinent literature by
Leschber (1969), whose arguments are not, in my opinion, convincing or
well founded. This author cited the findings of Bucksteeg and Thiele
(1957), which, as noted above, cannot be judged very reliable; he had
been especially impressed by the fact that a cuprocyanide, said to have
been KjCu^N)^ proved more toxic than K2Ni(CN)4- He believed the dis-
sociation of the nickelocyanide complex to be much greater than that of
the cuprocyanide complex. Although obviously well aware of the work of
Doudoroff, Leduc, and Schneider (1966), Leschber evidently overlooked
the fact, stressed by the latter authors, but overlooked also by Blaha
(1968), that the value of 10"^ for the cumulative dissociation constant
of the nickelocyanide ion has been repeatedly shown to be greater than
the true value by some 8 to 9 orders of magnitude. He probably, also,
seriously underestimated and overestimated amounts of free cyanide and
cuprous ion, respectively, present in dilute cuprocyanide solutions at
equilibrium when the CN/Cu mole ratio is as great as 4.0. Calculations
in which the stepwise dissociation of the cuprocyanide ions is correctly
dealt with show that in very dilute test solutions of the indicated
composition much of the cyanide but not of the copper must be free.
It is apparent that the stepwise dissociation of the cuprocyanide
95
-------
•Z fj
complexes, Cu(CN)4 and Cu(CN) ~*• t has not been properly considered in
any calculations of Leschber (19690 and also of Blaha (1968') and Stumm,
Woker, and Fischer (1954), only one dissociation constant having been
mentioned and considered as pertinent to the equilibrium in each case.
Upon completion of the studies considered above, a number of interesting
questions remained to be answered concerning the apparent toxicity of the
silver-cyanide complex ion and the influence of chloride ion on the
dissociation and toxicity of the complex, the possible toxicity of
cuprocyanide complex ions, the prolonged survival of fish in solutions
of the iron-cyanide complexes and recently prepared solutions of the
nickelocyanide complex with low pH whose computed equilibrium levels of
HCN should have proved rapidly fatal, and so forth. After some further
investigation of these matters by Doudoroff (unpublished), the task of
answering many of the more important remaining questions was undertaken
and successfully accomplished by Broderius (1973). The results of his
detailed studies of the chemistry and toxicology of metallocyanide
complexes are presented, together with various other pertinent data not
yet mentioned or adequately discussed, in the following summaries of
available information concerning the acute toxicity of various individual
metal-cyanide complexes or groups of similar or closely related complexes.
Almost no information on the chronic or sublethal toxicity of the cyanide
complexes has been found.
ZINC-CYANIDE AND CADMIUM-CYANIDE COMPLEXES
Doudoroff (1956) and Broderius (1973) have noted that the dissociation
constants of the order of 10~17 or 10~19 attributed to the zinc-cyanide
and cadmium-cyanide complex ions, Zn(CN)4~2 and Cd(CN)4~2, do not signify
slight dissociation of the complexes in dilute solutions even when the
pH is high. On the contrary, computations have shown that, in very
dilute but acutely toxic solutions of these complexes, with total cyanide
96
-------
concentrations less than 0.3 mg/1 as CN, the dissociation should be
almost total at any level of pH commonly encountered in natural waters.
Thus, it seems reasonable to conclude that waste cyanide complexed with
zinc or cadmium in concentrated, alkaline solutions (wastewaters) should
be about as dangerous as free cyanide to fish in receiving waters, if
not more dangerous.
Doudoroff (1956) found dilute solutions of the two complexes to be
apparently somewhat more toxic to fathead minnows than comparable NaCN
solutions with the same total cyanide content. The 96-hour median
tolerance limits of NaCN and of solutions of the zinc-cyanide and
cadmium-cyanide complexes prepared by combining NaCN with ZnSO^ or
CdSO^ and having cyanide/metal mole ratios slightly less than 4.0 were
found to be equivalent to about 0.23, 0.18, and 0.17 mg/1 of cyanide
as CN, respectively. The concentrations of the metals, especially of
zinc, in the solutions of the complexes that proved lethal to about 50
percent of the test animals were far below levels tolerated by most of
the fish for 96 hours in the absence of cyanide. These results indicate
cooperative joint action or synergism (in a broad sense, as explained
farther) of free cyanide and zinc or cadmium ion; however, they cannot
be regarded as a conclusive demonstration of such interaction, only one
toxicity bioassay of NaCN and of each of the two complex cyanide solu-
tions having been performed.
Chen and Selleck (1969) observed a similar interaction of zinc and
cyanide in experiments with guppies. They concluded that the "additive
nature of the zinc and cyanide toxicities is indicated clearly, i.e.,
the permissible zinc concentration increases with decreasing cyanide
concentration and vice versa." Unfortunately, they also stated that an
"antagonistic effect was noted between the zinc and cyanide ions because
a mixture of the two ions always has less toxicity than that obtained by
the simple addition of the toxicities of the two components determined
97
-------
separately under otherwise identical conditions." This statement
reflects not only a lack of recognition of the greater importance as a
toxicant of molecular HCN than of cyanide ion, but also, more impor-
tantly, an all too common misconception leading to misuse of the term
"antagonism", which denotes active opposition, resistance, or counter-
action.
The error involved in the system of terminology that equates or confuses
strictly additive joint toxicity of two or more combined toxicants
Cwhich is merely one kind or level of cooperative action) with a lack of
any interaction of the toxicants has been emphatically pointed out by
Sprague (1970) and by Warren (1971). As Sprague has explained, this
system, adopted by several authors in the past, "fails to cover certain
categories of effect adequately. In particular, it ignores the case
where each of two toxicants acts as if the other were not present. Such
a case ends up by default in the category 'antagonism'. Similarly, two
toxicants which have a combined effect somewhat greater than either one
separately, but less-than-additive, would be relegated to "antagonism1.
Obviously, the term antagonism, which signifies counteracting or opposing
effects, cannot include the case where two toxicants do not affect each
other's action nor the case where they work together somewhat."
Warren (1971), who has discussed the terminological problems in more
detail, has pointed out also that the term antagonism should be reserved
for physiological phenomena and not applied to chemical and physical
reactions that occur in the external medium of an organism. The reaction
of highly toxic heavy metal cations with free cyanide to form relatively
harmless complex ions or precipitates of insoluble metal cyanides in the
medium of aquatic organisms is not true antagonism. The effective toxi-
cants having been eliminated from the medium through chemical reaction,
they cannot be correctly said to counteract each other or to have
opposing effects on the organisms.
98
-------
Both Sprague C1970) and Warren (1971) prefer to avoid the use of the
term synergism because of its ambiguousness. In a broad sense and, in
my view, the most correct sense, synergism denotes truly joint or coop-
erative action at any level, including additive and infra-additive (less-
than-additive) interaction, with the exception only of "response addi-
tion" (Anderson, 1974) already explained. The term has also been often
used, however, in a narrower sense, as a synonym of potentiation,
supplemental synergism, or supra-additive (more-than-additive) joint
action. It has thus indeed lost much of its usefulness. But the term
antagonism also has become ambiguous because of its frequent misuse in
referring to infra-additive joint action that involves no counteraction.
In my opinion, both terms are still useful when employed correctly and
defined clearly, and so they are used in this review.
Even infra-additive synergism of cyanide and zinc and of cyanide and
cadmium has not always been observed. Doudoroff, Leduc, and Schneider
(1966) found the median survival times of bluegills in solutions of
cyanide (0.40-0.64 mg/1) combined with zinc or cadmium (one solution
with each metal) to be slightly greater than those predictable on the
basis of determined molecular HCN levels, the known toxicity of HCN to
the fish, and the assumption that the metals would not contribute to
the toxicity of the mixtures. The differences between the latter
median survival times and those experimentally determined, which were
not long (256 and 134 minutes), were too small to be interpreted as
indicating true antagonism of the combined toxicants. They may well
have been due simply to unavoidable experimental error, but they do not
suggest any synergism or cooperative action of the toxicants. The data
of Stumm, Woker, and Fischer (1954) also indicate some reduction of the
toxicity of cyanide in the presence of cadmium or zinc, but the concen-
trations of the two complexes tested by them (10-100 mg/1 as CN) were
much too high and too rapidly fatal for these results to be very
meaningful or instructive.
99
-------
Bucksteeg and Thiele (1957) reported minimum concentrations harmful to
fish of KCN, K2Zn(CN)4, and K2Cd(CN)4 to be 0.1, 0.3, and 0.75 mg/1 as
CN, respectively. These values indicate very pronounced reduction of
the toxicity of cyanide in the presence of the two heavy metals in
question, especially of cadmium. One can have no confidence in their
validity, however, in the absence of sufficient information about their
experimental derivation and in view of all the contradictory experi-
mental evidence and theoretical considerations presented above.
Doudoroff (1956), whose experiments included some toxicity tests of
purified K2Zn(CN)4, and others have shown that complexation of free
cyanide with zinc and cadmium certainly does not render it so much less
dangerous to fish in very dilute solutions.
Cairns and Scheier (1968) reported some reduction of the toxicity of
cyanide to bluegills when KCN was combined with zinc chloride, ZnCl2,
in dilute test solutions. The 96-hour median tolerance limits reported
were 0.18 mg/1 as CN for KCN alone and 0.26 mg/1 as CN for the mixture.
The two salts were combined in such proportions that, when the concen-
tration of cyanide (total) was 0.26 mg/1 as CN, that of zinc was 3.9
mg/1, a level only slightly below the 96-hour median tolerance limit of
the metal ion, determined by testing solutions of ZnCl2 alone and
reported to have been 4.2 mg/1. The mole ratio of zinc to cyanide
2
obviously was far in excess of 0.25, the ratio for the Zn(CN)4~
complex. The authors suggested that the apparent reduction of the
toxicity of cyanide in the presence of the added zinc was due probably
to complexation. However, there is no evidence that they undertook any
calculations to determine how much of the cyanide remained in solution,
or that none was present as a zinc cyanide precipitate (i.e., that the
solubility product constant for Zn(CN)2 was not exceeded), and then to
evaluate the extent of complexation of cyanide and zinc that could have
occurred in their test solutions. I have not attempted such calcula-
tions, which would have been quite complicated and would have required
100
-------
knowledge of the pH of the test solutions (which was not reported) or
the use of assumed values, but I am almost certain that no material
complexation of zinc and cyanide was possible at any possible pH of the
solutions. The reported toxicity test result may have been due simply
to experimental error.
It can be concluded that there is more impressive evidence of synergism
of cyanide and zinc or cadmium in very dilute but still acutely toxic
solutions of the complexes in question than of their true antagonism,
but no interaction of great importance has been demonstrated. Complex-
ation of cyanide with the two metals certainly can be disregarded in
establishing maximum permissible concentrations of these toxicants when
they are present together in receiving waters in which fish are to be
protected. It clearly cannot be important in any solution with a level
of free cyanide that can be regarded as safe for fish life, and the
contrary finding of Bucksteeg and Thiele (1957) must be discounted.
It can be noted, additionally, that Doudoroff (1956) showed that the
introduction into a fairly concentrated, clear solution of the zinc-
cyanide complex of enough NiSO* to produce a small amount of precipitate
(by replacement of some of the zinc with nickel in the complex) did not
materially reduce the great toxicity of the solution to fathead minnows.
This result was to be expected, since most of the cyanide in the solu-
tion remained in the form of zinc-cyanide complex after the addition of
nickel, and the dissociation of this complex upon dilution of such a
solution could not be affected by the presence of a small amount of
nickel.
NICKEL-CYANIDE COMPLEX
Milne (1950) found that goldfish were not noticeably affected after
exposure for 24 hours to a solution in which NaCN and NiS04 were combined
101
-------
to form the nickelocyanide complex and whose pH. was 8.1. The measured
total cyanide content of this test solution was 104 mg/1 and the nickel
content 77 mg/1, a very excessive amount if both reported values are
correct, for they indicate a CN/Ni mole ratio of 3.05 and not near 4.0.
The pertinent findings of Stumm, Woker, and Fischer (1954) and of
Bucksteeg and Thiele (1961), already discussed, also indicate low or
only moderate toxicity to fish of the nickelocyanide complex but are not
highly instructive.
Doudoroff (1956) tested for toxicity to fathead minnows at 20° C some
nickelocyanide solutions prepared by slowly adding a fairly concentrated
solution of NiS04 in distilled water to a like, agitated solution of
NaCN until a small amount of Ni(CN)2 precipitate appeared. The CN/Ni
mole ratio was either 3.95 or 3.81 (in one early experiment only). The
solutions prepared in the above manner, which were quite alkaline in
reaction, were diluted as necessary for the toxicity tests with a
synthetic, soft water (total alkalinity 17.5 mg/1 as CaCOjj total hard-
ness 25 mg/1 as CaCO^). When the dilutions were neither renewed
periodically nor artificially aerated, but remained adequately oxygenated
through absorption of oxygen at the surface exposed to the atmosphere
(dissolved oxygen usually 5 to 6 mg/1 after the first 48 hours), the
nickelocyanide complex eventually proved highly toxic to the fish. Some
of the fish (20-40%) died within 168 hours at a total cyanide concentra-
tion of 0.5 mg/1 as CN, and the minimum concentrations necessary to kill
half of the fish in 96 and 168 hours were estimated to be 0.95 and 0.65
mg/1 as CN, respectively. However, at no tested concentration below
600 mg/1 as CN were all of the fish, or more than 80 percent, killed
within 24 hours. At concentrations of 8 to 67 mg/1 as CN, all of the
fish died in less than 48 hours, but at the concentration of 200 mg/1
as CN, all lived for 24 hours and 80 to 90 percent survived for 48 hours.
The result of a single test of a purified potassium nickelocyanide,
, at the concentration of 2.0 mg/1 as CN agreed very well with
102
-------
those of comparable tests of the mixture of NaCN and NiS04 solutions
diluted so that the total cyanide level was the same.
It was soon determined that the toxicity of all the initially more or
less alkaline, standing test solutions increased markedly with time
after introduction of the test animals. This increase of toxicity was
shown to be due mostly to the decrease of the pH of the solutions caused
by the production of carbon dioxide, C02, by the respiring fish.
Periodic renewal of test solutions (i.e., their replacement with fresh
solutions to which the fish were transferred), or mere addition to them
from time to time of small amounts of sodium hydroxide, NaOH, so as to
return the pH approximately to its initial value, greatly prolonged the
survival of the fish, or made possible their indefinite survival, at a
concentration of the complex that had proved fatal within two days with-
out such renewal or treatment of the solutions. An increase of the pH
of an aged test solution by only about half a pH unit upon addition of
NaOH was found to be sufficient to bring about prompt recovery of fish
that had already been visibly affected (i.e., in distress) in the same
medium before the treatment. Dilute solutions proved more rapidly fatal
than much more concentrated ones evidently because critical pH values
were attained sooner in the former solutions than in the more concentra-
ted and initially more alkaline solutions as the pH of all was reduced
by the respiration of the fish.
In a subsequent experiment, Doudoroff used five different waters of
varying total alkalinity (synthetic waters resembling natural waters with
total alkalinity values of 5 to 192 mg/1 as CaC03) to dilute a concentra-
ted stock solution in which 5,150 mg/1 of cyanide as CN had been combined
with 2,940 mg/1 of nickel (CN/Ni mole ratio 3.95). The pH values and
dissolved oxygen concentrations of the dilution waters were reduced to
103
-------
nearly constant levels by holding test animals in jars (test vessels)
containing these waters for 2 days before any toxicant was added. The
pH was promptly readjusted to these levels, as necessary, by addition
of €09 after the stock niekelocyanide solution had been added to these
waters to make up test solutions of the various desired concentrations.
Thus, toxicity tests were performed and the 24-, 48-, and 96-hour median
tolerance limits (TL^'s) were determined at each of five different,
nearly constant levels of pH ranging from about 6.5 to 8.0 (mean values)
and at nearly uniform oxygen concentrations. The 24-hour TL^Q values so
determined, when expressed as initial total cyanide (CN) concentrations,
ranged from 1.35 mg/1 Cat pH 6.5) to 1,300 mg/1 (at pH 8.0); the 48-
hour TL50 values ranged from 0.75 mg/1 (at pH 6.5-6.6) to 1,200 mg/1
(at pH 8.0); and the corresponding 96-hour values ranged from 0.42 to
730 mg/1. Each curve relating the logarithms of the TLgQ values for a
given exposure period to the mean pH values was sigmoid in shape. The
largest increases of log TL^Q with a given increase of pH occurred in
the pH range of 7.4 or 7.5 to 7.8. With increase of pH from 7.5 to 7.8,
the TLgQ values increased about 10-fold to 13-fold. The magnitude of
these increases agrees fairly well with that of the (roughly) 16-fold
increase in concentration of the nickelocyanide complex in slightly
alkaline solutions that is required, according to theory as explained
earlier in this review, for maintaining a constant concentration of
molecular HCN when the pH is increased by three tenths of a pH unit.
In view of the approximate agreement just mentioned of the maximum
slopes of theoretical and experimentally derived curves, Doudoroff
concluded that the acute toxicity of dilute and not very alkaline
solutions of sodium nickelocyanide must be determined chiefly by their
molecular HCN content. The striking influence of pH on the toxicity
could not be explained otherwise. However, in the year 1956, it was
not possible to reconcile the available toxicity data and the proposed
explanation thereof with the then generally accepted value for the
104
-------
cumulative dissociation or instability constant of the Ni(CN)4~2 complex
22
ion, namely, 10~ . Having made the necessary calculations, and assum-
ing that the toxicity of molecular HCN is not materially reduced in the
presence of ionic nickel, Doudoroff concluded that this value for the
constant, KD, probably is grossly erroneous and that a value of the
order of 10~30 would better accord with the toxicity test results. He
noted that the tested solutions should have had a stronger odor of HCN
than they did, as well as much greater toxicity to fish, if the correct
value for the constant were nearly as great as 10~22. Freund and
Schneider (1959) thereupon redetermined the constant at 25° C, finding
it to be about 1 x 10~31, but Schneider and Freund (1962) later recom-
puted it, using a new value for the ionization constant of HCN that was
deemed most reliable, and reported a corrected value for KD of 5.4 x
10" . Results in fairly close agreement with these findings and with
Doudoroff's (1956) estimate were subsequently reported by Christensen
e£ al. (1963) (indicated KD 7.9 x 10~31 at 25° C) and by Broderius
(1973), whose somewhat variable "apparent KJJ" values based on determined
HCN levels in various nickelocyanide solutions prepared by him averaged
1.0 x 10~31.
Complete agreement of Doudoroff's (1956) toxicity data with calculated
molecular HCN concentrations in the test solutions at equilibrium
evidently could not be achieved by using the correct dissociation con-
stant for the Ni(CN)4~2 ion in the calculations. Outside the pH range
of about 7.4 to 7.8, serious discrepancies remained that needed to be
explained. The toxicity of relatively concentrated solutions with pH
near 8.0 was apparently greater than that which could have been due
entirely to their free cyanide or molecular HCN content. On the other
hand, the observed toxicity of very dilute solutions that were slightly
acid or nearly neutral was much less than that which could have been
expected on the basis of the correctly computed equilibrium concentra-
tions of molecular HCN and the assumption that the low-level nickel ion
105
-------
present exerted no material, truly antagonistic effect counteracting the
toxic action of HCN.
Doudoroff (1956) observed that some of the ions present in relative
abundance in his most concentrated and alkaline solutions could well
— *?
have contributed materially to their toxicity. The NiCCN)^ ^ ion
certainly could not be supposed to be entirely harmless at very high
concentrations. As for the fact that the dilute solutions that were
neutral or slightly acid proved less toxic than they apparently should
have been, Doudoroff was unable in the year 1956 to propose an adequate
explanation of it; his explanation was only a partial one. He pointed
out that when the fraction of the total cyanide that is free cyanide
becomes large, the difference between total cyanide and cyanide bound in
complex ions becomes important, i.e., it can no longer be neglected in
considering or predicting effects of further reductions of pH on the
toxicity of the complex. He further noted that, in the absence of a
large reservoir of potential toxicant in the form of complex ions,
gradual loss of cyanide from very dilute, acid and neutral test solutions
during extended bioassays could have been considerable, augmenting the
initial concentrations necessary to render the solutions lethal to fish.
But the possibility of important effects of other, unidentified factors
also was recognized by him. Only much later did he realize, as reported
by Broderius (1973), that equilibria in his more dilute test solutions
with relatively low pH may not have been attained for a long time after
the tests of the fresh media were begun. He then found that when such
solutions were aged by storing them in sealed vessels (without fish),
their toxicity increased markedly, approaching the levels expected on
the basis of theoretical considerations upon attainment of equilibria.
Thereafter, Broderius (1973) undertook many additional experiments
designed further to elucidate the significance of the observations
reported above.
106
-------
Broderius' (1973) findings concerning the rates of dissociation and
f\
formation of the NiCCN)^"^ complex ion in solutions of varying pH and
total cyanide content already have been summarized briefly. Having
determined how much time was required for attainment of equilibria in
dilutions of a concentrated nickelocyanide solution, he was able to
evaluate both the molecular HCN content and the toxicity to bluegills
of dilutions in which equilibria had been nearly or fully attained.
When the total cyanide concentrations in the test solutions, with pH
7.1 or 6.5, did not exceed 10 mg/1 as CN and the average HCN concentra-
tion was 0.20 mg/1, the median survival times of the bluegills (240-257
minutes) agreed well with that of fish exposed to an NaCN solution
(pH 7.1) with the same determined HCN content (260 minutes). The solu-
tion with pH 6.5 contained initially 0.23 mg/1 of cyanide combined with
nickel and had been aged for 20 days before it was tested, when the
dissociation of the complex ions was believed to have been almost total;
a small amount of cyanide initially present in this solution and about
the same amount initially present in the NaCN solution apparently were
lost during the tests. Nickelocyanide solutions exactly like the last
mentioned one (pH 6.5) but aged for only 5 and 8 days had determined HCN
levels of only 0.08 and 0.11 mg/1, respectively, and killed no bluegills
within a period of 480 minutes, after which the tests were discontinued.
The median survival times of bluegills in similar, sufficiently aged
(for 1 day) nickelocyanide solutions with determined molecular HCN
content 0.20 mg/1 but with total cyanide concentrations of 248 and 500
mg/1 as CN and with pH values of 7.45 and 7.60 were only 196 and 140
minutes, respectively. A linear relation of median survival time to
total cyanide concentration in nickelocyanide solutions with varying pH
but a constant molecular HCN concentration is indicated. The indicated,
moderate increase of toxicity associated with great increase of total
cyanide in solutions of constant HCN content (but different pH) is of
107
-------
little practical importance but is nevertheless noteworthy. It clearly
was too pronounced to be ascribed to the increase of the always very low
concentration of cyanide ion accompanying the increase of pH to levels
not exceeding 7.6, and it could most reasonably be attributed to
toxicity and synergistic action of the nickelocyanide ion itself.
Evidence of the ability of the nickelocyanide complex ion to penetrate
the tissues of fish exposed to toxic nickelocyanide solutions has been
presented by Broderius. He used C-labeled cyanide combined with
nickel, and he compared the rates of accumulation of the radioactive
carbon in tissues of bluegills exposed to solutions containing the
labeled complex and to solutions of NaCN alone, also containing
C-labeled cyanide. The concentrations of molecular HCN in the
nickelocyanide solutions were approximately the same as those in the
NaCN solutions. Fairly reliable estimates thus could be made of the
rates of accumulation of ^C in the tissues of fish exposed to the
former solutions that was not a consequence of, or referable to, the
o
intake of HCN. Broderius concluded that the NiCCN)^ complex ion does
not enter the fish body readily, but its limited penetration of the
tissues of fish exposed to relatively concentrated and alkaline
nickelocyanide solutions may well be the reason why the toxicity of
these solutions did not appear to be attributable entirely to their
molecular HCN content. The *x! accumulated in the gill tissues much
more markedly than it did in the blood and in tissues of internal organs
sampled.
The experiments of Doudoroff, Leduc, and Schneider (1966), whose methods
and findings already have been described, did not reveal any considerable
influence of the Ni(CN)4~2 ion at concentrations up to 500 mg/1 as CN on
the acute toxicity of nickelocyanide solutions. These experiments were
quite similar to those of Broderius that led him to the conclusion that
108
-------
the complex ion itself may have considerable toxicity, effective even in
tests of short duration, at concentrations of 500 mg/1 as CN and less.
No explanation can be offered for the disagreement of the results in
question, and I cannot say whose evidence is more reliable. Broderius'
methods were refined but his pertinent tests fewer than those of
Doudoroff, Leduc, and Schneider. Even if Broderius1 finding is correct,
it certainly does not invalidate the conclusion of Doudoroff, Leduc, and
Schneider that reduction of the pH of nickelocyanide solutions (test
media) at gill surfaces by carbon dioxide released there did not cause
the HCN concentrations there to be materially greater than those in the
inspired, ambient media. This conclusion was based mostly on a compari-
son of median survival times of bluegills exposed to poorly and well
buffered, neutral or slightly alkaline nickelocyanide solutions (100 mg/1
as CN) of widely varying free CC^ content, whose variations were not
found to have any appreciable effect on the toxicity of solutions of
uniform pH. The reasoning and computations that led to this conclusion,
which is not deemed of major importance, are too involved to be fully
or adequately reported here with the supporting data. The conclusion,
which implies that the dissociation of Ni(CN)4~^ ions upon reduction of
the pH of a solution cannot be very rapid, is now further supported by
Broderius' data on the rates of dissociation of the complex ions. It
is apparent that liberation of 14C-labeled cyanide at the gill surfaces
was not the reason why some of the labeled cyanide present in test
solutions in the form of nickelocyanide ions, and not as molecular HCN,
was evidently able to penetrate the tissues of bluegills in the experi-
ments reported by Broderius. Penetration of the gills by the complex
ions themselves apparently did occur.
Doudoroff (1956) observed that the bodies of fathead minnows exposed to
fairly concentrated nickelocyanide solutions that were alkaline enough
not to be rapidly lethal to the fish became often, but not invariably or
with any apparent regularity, exceedingly swollen some time before the
109
-------
fish died, perhaps of osmoregulatory failure. Red blotches, apparently
due to internal hemorrhages, commonly appeared on the abdomens of the
fish so affected. No such signs of intoxication were observed in any
fish dying, apparently of HCN poisoning mainly, in test solutions that
were more dilute than the above-mentioned ones but were about as toxic
because of their lower pH. A distinctive kind of subacute poisoning
with penetrating nickelocyanide ion of the fish exposed to high concen-
trations of the ion thus is suggested by the reported, curious observa-
tions .
The practical significance of the reviewed information on the toxicity
of the nickelocyanide complex, whose acute toxicity to fish has been
studied more intensively than that of any other metal-cyanide complex,
will now be considered. Complexation of free cyanide with nickel obvi-
ously can be very effective in protecting fish from lethal effects of
the cyanide in alkaline waters (e.g., at pH near 7.5), but not at all
effective in decidedly acid waters Cat pH 6.5 or less) in which HCN
levels near equilibrium levels can be attained. In very dilute solutions
with pH near 6.5 or less, the dissociation of the complex that tends to
render the solutions acutely toxic is very slow, however; less than one
third of it may take place during the first day. If the cyanide gradu-
ally liberated at the low pH in the form of volatile HCN is rapidly lost
to the atmosphere or otherwise, as it may very well be lost from even
moderately turbulent waters of receiving streams, even a fairly high
initial concentration of the complex may never become fatal to fish.
But let us assume that virtually no liberated cyanide will be lost to
the atmosphere or otherwise eliminated during the time period required
for attainment of equilibrium. In that case, any dilution of the complex
with so low a pH (6.5 or less) and with a cyanide content great enough
to kill sensitive fishes when the cyanide is all in the free state will
eventually become lethal to the fish. The dissociation of the complex
ion will be virtually total, or, if the initial concentration of the
110
-------
complex and the pH are not low enough for such complete dissociation to
take place, the dissociation will be more than sufficient to render the
water acutely toxic to fish. In alkaline waters, on the other hand,
total cyanide levels many times the minimum lethal concentration of
free cyanide may be necessary to render the waters acutely toxic, even .
when no cyanide is lost.
We must also consider, however, just how well complexation of free cya-
nide with nickel can protect sensitive fishes in alkaline waters against
the harmful effects of extremely low, sublethal concentrations of free
cyanide. My calculations show that it can have very little effect or
be wholly ineffective even when the pH of the water receiving the complex
remains at a level as high as 8.0, if there is no rapid loss of cyanide.
For example, let us assume that the highest concentration of molecular
HCN that would have no serious adverse effect on a sensitive species of
fish inhabiting a given water with the pH of 8.0 is only 2 x 10"' moles
per liter, or about 0.005 mg/1. We have seen that as little as 0.01 mg/1
of free cyanide (mostly HCN) has a speedy, very pronounced, and lasting
effect on the swimming ability of salmonid fishes; thus, concentrations
even well below 0.005 mg/1 may be shown by further research to be decid-
edly harmful to these fish. The latter concentration has been considered
or proposed by some authorities as a reasonable upper limit for permissi-
ble free cyanide concentrations in waters in which valuable fishes are to
be well protected. Substituting the appropriate values Grounded approxi-
mations) in the pertinent ionic equilibrium equations already presented
and explained, after computing (solving for) the values of [CN~] and
[Ni(CN)4"2], we get:
111
-------
[H+] [OH [1(T8] ClO'8]
[HCN] [2 x 10'7]
[Ni+2] [CM']4 [5 x 1(T8] [10-8]4
[Ni(CN)4-2]
= 5 x 10~10 = Ka, and
= 5 x 10"31 = Kn,
since the ionization of HCN at pH 8.0 is slight, and so [Ni+2], which
should be equal to ([HCN] + [CN~])/4, can be taken to be simply equal to
[HCN]/4. It can be seen from the first one of these equations that when
the concentration of molecular HCN is 2 x 10~7 moles per liter (0.0052
mg/1 as CN) and the pH is 8.0, or [H+] is 10 , the concentration of the
CN~ ion must be about 10~8 also (0.00026 mg/1). From the second equation,
it can be seen that, if all of the free cyanide (HCN and CN~) derives
from dissociation of the Ni(CN)^~2 complex ion, and the value for the
dissociation constant (Kp) of this ion that has been used above
(5 x lO"3*) is very nearly correct and applicable, then the concentra-
tion of the undissociated Ni(CN)^ ion must be about 1 x 10~" moles per
liter, or equivalent to about 4 x 10 moles of cyanide (CN) per liter
(0.0001 mg/1). Thus, of the total cyanide (about 0.0056 mg/1 as CN) ,
less than 2 percent must be present in the form of the complex ion at
equilibrium under the specified conditions, the remainder being present
as free cyanide, and most of this (nearly 95%) as molecular HCN. In
other words, the dissociation of the complex must be almost total.
If the cumulative dissociation constant of the tetracyanonickelate(II)
ion is now taken to be as low as 1 x 10~3 , the average of "apparent Kn"
values computed by Broderius (1973) from determined HCN concentrations
in his various nickelocyanide solutions at 20° C, the result of the
calculation in question is not seriously altered. Computations then
show that about 9 percent of the total cyanide should be bound in com-
plex ions at pH 8.0, but only about 1 percent at pH 7.7, when the HCN
112
-------
concentration is 2 x 10~7 moles per liter. The pH values of most fresh,
surface waters receiving waste cyanides are not much above 8.0. Com-
plexation of free cyanide with nickel before discharge evidently cannot
be expected to prevent the occurrence in these waters of free cyanide
levels above 0.005 mg/1 when the levels of total cyanide introduced in
the form of the complex considerably exceed 0.005 mg/1 and persist until
equilibria are fully or nearly attained. The addition of nickel ion to
moderately alkaline waters seriously polluted with free cyanide obviously
also is not likely to render these waters entirely suitable for sensitive
fishes, though it often could reduce the free cyanide concentrations to
levels that these fishes can tolerate. It is appropriate to remark here
that, when the possible effectiveness of complexation of cyanide with
nickel as a means of preventing the occurrence of lethal levels of free
cyanide in nearly neutral waters is evaluated,, the choice of values to
be used in calculations for the dissociation constant of the complex and
the HCN tolerance limit of fish to be protected becomes critically
important. Use of different values whose differences may seem to be
minor and probably unimportant can lead to very different conclusions.
SILVER-CYANIDE COMPLEX
Doudoroff, Leduc, and Schneider (1966) observed that the toxicity to
bluegills of a somewhat alkaline solution (pH 7.5) containing 10 mg/1 of
cyanide (CN) complexed with silver was not much greater than that of a
similar but slightly acid solution (pH 6.5) with the same total cyanide
content. The determined molecular HCN concentration in the former solu-
tion was only 0.02 mg/1, whereas the concentration in the slightly acid
solution was 0.12 mg/1. The median resistance times of the bluegills
exposed to the two solutions were nearly the same (833 and 789 minutes),
and so were evidently almost independent of the HCN concentrations. A
similar solution with the same total cyanide content (10 mg/1 as CN) but
a pH of 6.0 and a determined molecular HCN content of 0.19 mg/1 was much
113
-------
more rapidly fatal, and its greater toxicity was evidently attributable
entirely or almost entirely to the high HCN concentration. In the solu-
tions with higher pH, the dying fish often showed no signs of cyanide
poisoning. Pronounced, superficial coagulation of mucus suggestive of
heavy-metal poisoning was usually observed in fish that had been exposed
for some time before death to the solutions containing the silver-cyanide
complex. These observations indicated that the AgCCN)2~ ion itself,
unlike the Ni(CN)4~2 ion, has fairly high toxicity to fish. The possi-
bility that the extremely toxic silver ion present in small amounts
because of dissociation of the complex caused or contributed to the
toxicity of some of the tested solutions (at pH 6.5-7.5) was considered,
but an important role of this toxicant was not demonstrated. Also
considered was the probability of increased dissociation of the complex
and production of HCN because of combination of the silver ion with
chloride ion, which was fairly abundant in the acid test solutions
because these had been prepared with water acidified with HC1.
Broderius (1973) explored more fully the toxicity of the silver-cyanide
complex in solutions of varying salinity and the influence of chloride
ion on the dissociation of the complex. The rapid dissociation of the
Ag(CN)2~ ion in fresh water was found to be somewhat more extensive at
equilibrium than that which was to be expected on the basis of calcula-
tions that could be made, using any one of the recently published values
for the dissociation constant of the ion. As in the case of the
nickelocyanide complex, the apparent dissociation constants for the
Ag(CN)2~ ion computed from determined HCN concentrations in solutions of
varying total cyanide content and pH, all at 20° C, varied considerably,
perhaps largely because of differences in ionic strength of the solutions.
Values computed from results of tests of solutions with total cyanide
levels of 0.5 to 20 mg/1 (pH 6.0-8.5) and 100 to 200 mg/1 (pH 7.1-7.5)
averaged 0.86 x 10~19 and 8.4 x 10"19, respectively. The over-all
average of the 14 computed values, pertaining mostly to total cyanide
114
-------
concentrations of about 10 mg/1 (5-20 mg/1), was determined to be 1.94 x
10" and by some 1 to 3 orders of magnitude greater than the recently
published values for the constant.
Even when a value as high as 2 x 10~19 is used as the dissociation con-
stant for the Ag(CN)2~ i°n i-n calculating degrees of dissociation in
very dilute solutions, complexation of cyanide with silver can be seen
to be quite effective in preventing the occurrence in cyanide-
contaminated fresh waters of sublethal levels of free cyanide believed
to be harmful to fish. An example of such a calculation, demonstrating
the relative ineffectiveness of complexation with nickel, already has
been presented. My similar calculation, in which 2 x 10~19 was taken
to be the correct value for KD, showed that, when the molecular HCN or
free cyanide level in a solution in which cyanide is complexed with
silver is only 0.005 mg/1 (2 x 10"^ M) , most of the total cyanide present
(more than 80 percent) must still be bound in the complex ions even at a
pH as low as 7.0. At pH 8.0, the dissociation is shown to be very
slight, a concentration of the complex (5 x 10 M) equivalent to about
2.6 mg/1 of cyanide as CN apparently being required to produce a molecular
HCN level of 0.005 mg/1.
Fresh water and sea water diluted to several different levels of
chlorinity (concentration of chloride ion, Cl~) were used by Broderius
(1973) as diluents in preparing test solutions containing 10 mg/1 of
cyanide (as CN) complexed with silver and having pH values near 7.7.
The molecular HCN concentrations in these solutions were found to increase
linearly from less than 0.01 to about 0.24 mg/1 with increase of the
chlorinity from nearly 0 to about 8.6 parts per thousand (ppt), the
chlorinity of a 50 percent dilution of the sea water used. With further
increase of the chlorinity to about 17.2 ppt (the chlorinity of the sea
water), the HCN concentration did not continue to increase, remaining at
the level of about 0.24 mg/1. The median survival time of euryhaline
115
-------
threespine sticklebacks in a solution prepared with water that was 10
percent sea water (chlorinity less than 2 ppt) was 786 minutes and nearly
the same as it was in the fresh-water solution (770 minutes), although
the HCN content had increased to about 0.05 mg/1. With further increase
of the chlorinity to that of the full-strength sea water, however, the
median survival time declined progressively in a curvilinear fashion; it
was 140 minutes at the chlorinity of the full-strength sea water and 260
minutes at the chlorinity one-half as great. These results show that
the toxicity of the solutions of low chlorinity was independent of the
very low concentrations of molecular HCN and silver ion, having been
determined by the virtually constant concentration of the moderately
toxic Ag(CN)2~ ion. The products of the dissociation of the complex
must have been largely or entirely responsible for the speedier lethal
action of the more saline solutions. The high HCN concentrations in the
solutions prepared with the more saline waters (50 to 100 percent but
not 25 percent sea water) apparently could have alone caused the death
of the fish in these solutions. The median survival times of the
sticklebacks in these saline solutions were approximately those that
were to be expected on the basis of results of some tests of comparable
NaCN solutions with the same chlorinities and not much different
molecular HCN levels. The latter tests, the results of which already
have been reported, revealed an influence of water salinity on the
toxicity of free cyanide or HCN quite similar to its effect on the
toxicity of the solutions of cyanide complexed with silver that had a
high and uniform determined HCN content. The toxicity of the Ag(CN)2~
ion may increase with increase of water salinity, as did the toxicity
of free cyanide, but such an effect is not clearly indicated.
The 24-hour median tolerance limits of the silver-cyanide complex for
sticklebacks in fresh water and sea water of about 17 ppt salinity were
found by Broderius to be approximately 6 and 3 mg/1, respectively, at
pH 7.7 and 7.9 and at a temperature of 20° C. The concentrations of
molecular HCN in the test solutions prepared with sea water were below
116
-------
0.1 mg/1, but this toxicant may have been at least partly responsible
for the acute toxicity of the solutions. In the fresh-water solutions
(pH 7.7), the HCN concentrations were well below 0.01 mg/1 and hardly
could have contributed materially to the measured, acute toxicity of the
solutions.
Bluegills proved more resistant to the Ag(CN)2~ ion in fresh water than
were the sticklebacks. The 24-hour median tolerance limit for the blue-
gills at pH 7.5-8.5 and 20° C evidently was not far above 10 mg/1 as CN,
the median survival times (for 10 fish) having been 29 and 31 hours at
this concentration of the complex in solutions with pH 7.5 and 8.5 and
determined HCN concentrations of 0.0123 and 0.0032 mg/1, respectively.
Perhaps because of difference in source or physiological condition of
the test animals, these results are not in very good agreement, however,
with those of Doudoroff, Leduc, and Schneider (1966), who recorded a
median survival time of bluegills of only about 14 hours in a solution
containing 10 mg/1 of cyanide (as CN) complexed with silver, at pH 7.5
and an HCN concentration of about 0.02 mg/1 (perhaps less). On the
other hand, Broderius' results fully support the view that the products
of dissociation of the silver-cyanide complex cannot contribute materi-
ally to the acute toxicity of solutions prepared with fresh water and
having pH values near or above 7.5. It can be seen that the median
survival time of the bluegills in the presence of 10 mg/1 of the complex
as CN decreased very little with a decrease of pH from 8.5 to 7.5 and a
nearly fourfold increase of HCN concentration from the extremely low
level of 0.0032 mg/1, which obviously could have had no appreciable
effect on the toxicity of the solution. Broderius als$ reported that
the median survival time of bluegills at a concentration of the silver-
cyanide complex of 7 mg/1 as CN, pH 8.0, and a molecular HCN level of
0.0068 mg/1 was 87 hours. A 96-hour median tolerance limit of the
Ag(CN)2~ ion somewhat below 7 mg/1 as CN is indicated.
117
-------
Broderius noted that the reaction of silver ion with chloride ion present
in sufficiently high concentration yields complex silver-chloride ions,
2
AgCl2~ and AgClj , as well as very slightly soluble silver chloride,
AgCl, with which chloride ions combine to form the complexes before any
AgCl can precipitate. By calculation, he found that the Ag(CN)2~ ion
must be the predominant silver complex species in a solution with a
CN/Ag mole ratio of 2.0 and containing chloride ion, which competes with
the cyanide ion for the silver ion. As the Cl~ ion concentration
increases from a low level, the dissociation of the Ag(CN)2~ complex ion
with liberation of CN~ ion also increases because of the intensified
competition by the Cl~ ion for the Ag"1" ion, but the extent to which the
cyanide complex can be thus decomposed must be quite limited, the
equilibrium level of free cyanide approaching an asymptote when most of
this complex is still intact. The observation that, in the solutions
prepared with diluted sea water and containing 10 mg/1 of cyanide (as
CN) initially complexed with silver, the HCN concentration did not
increase appreciably beyond the level of about 0.24 mg/1 with increase
of chlorinity beyond 8.6 ppt at a constant pH of 7.7 is in agreement
with that conclusion. It is noteworthy, however, that when the
chlorinity of a solution containing only 1 mg/1 of cyanide complexed
with silver was increased by the addition of sodium chloride, NaCl, the
HCN concentration continued to increase markedly, at pH 7.7, with
increase of chlorinity beyond 0.25 M, or 8.9 ppt, the chlorinity of
about a 50 percent dilution of sea water. It attained a level of 0.052
mg/1 at a chlorinity of 20.8 ppt, but was only 0.028 mg/1 at the
chlorinity of 8.9 ppt. The relatively low concentration of the cyanide
complex in the NaCl solutions may be the main reason why a maximum HCN
level was not nearly attained at the lower chlorinity; similar tests
with higher concentrations of the complex and with as much NaCl added
were not performed by Broderius.
118
-------
Evidence of penetration of the silver-cyanide complex ion into the bodies
of bluegills has been presented by Broderius, who studied the rates of
accumulation of silver in various tissues of fish exposed for brief
periods to solutions containing the complex or only AgNOg, all with a
pH of 8.0. Although the silver ion was found to penetrate quite readily
into the bodies of the fish, its concentrations in the tested solutions
in which the silver was complexed with cyanide were found by calculation
to have been too low for the entry of this ion to have been responsible
for much of the observed accumulation of silver in the internal tissues
of fish exposed to these solutions. When the fish were exposed to AgN03
solutions, silver tended to accumulate most markedly in gill tissues,
whereas fish exposed to solutions in which the silver was complexed with
cyanide accumulated little silver in the gill tissues, and accumulation
in their internal organs and blood was more pronounced than accumulation
in the gills. The silver-cyanide complex ion evidently passes through
the gills and into internal organs of fish more readily than does the
nickelocyanide complex, but not quite as readily as does the cuprocya-
nide complex ion CuCCN)2~ (to be considered presently), whose more rapid
penetration may be the reason for its greater toxicity.
Silver ion is known to be extremely toxic to fish (one of the most toxic
ions), and its concentration, in mg/1, must be about twice the HCN con-
centration in a fresh-water solution in which the CN/Ag mole ratio is
2.0, if there is no withdrawal of either silver ion or cyanide ion from
the system except by the complexation of silver with cyanide. It is
quite possible, therefore, that at very low concentrations of the silver-
cyanide complex and low levels of pH at which the concentration of
neither the complex ion nor of speedily acting HCN is high enough to be
soon fatal to fish, the toxicity of the free silver ion is often pre-
dominant. The silver ion under these circumstances could well be the
toxicant principally or entirely responsible for slow death of fish, or
cause chronic, sublethal injury to them more serious than the effects of
119
-------
the low concentrations of HCN and of the Ag(CN)2~ ion. The chronic
toxicity of metallocyanide complexes has not yet been investigated.
COPPER-CYANIDE COMPLEXES
In early experiments on the toxicity of the cuprocyanide complexes
reported by Stumm, Woker, and Fischer (1954), Lipschuetz and Copper
(1955), and Bucksteeg and Thiele (1957), the CN/Cu mole ratio ranged
from 3.0 to 4.0. The amounts of the cyanide group in the test solutions
thus greatly exceeded the minimum amounts necessary for combination with
all of the copper(I) present to form the dicyanocuprate(I) complex ion,
Cu(CN)2~- Marked reduction of the toxicity of cyanide upon its complex-
ation with copper(I) was demonstrated, but all of the tested combina-
tions of cyanide and copper proved quite toxic, even though the solutions
tested by Stumm, Woker, and Fischer, with a reported CN/Cu ratio of 3.0,
were very alkaline (pH 9.0-9.4).
The most detailed of the early studies apparently was that of Lipschuetz
and Cooper (1955), whose test animal was the western blacknose dace,
Rhinichthys atratulus meleagris. They found that the 24-hour median
tolerance limits for this fish of KCN alone and of KCN combined with
cuprous cyanide, CuCN, to produce cuprocyanide solutions with CN/Cu mole
ratios of 4.0, 3.75, and 3.0 were, respectively, about 0.22, 0.38, 0.47,
and 0.71 mg/1 as CN. It is evident that the toxicity of solutions of
equal total cyanide content decreased considerably with increase of the
amount of copper present, which must have resulted in decrease of the
concentration of free cyanide. The pH of the test solutions ranged from
7.6 to 8.0, and the temperatures from 20 to 22° C. The test solutions
were not aged to ensure attainment of equilibria, and, judging by the
data of Broderius (1973) on the rates of formation of cuprocyanide
complexes, one has good reason to doubt that equilibria had been nearly
attained. Other investigators presumably have dissolved or have diluted
120
-------
already formed cuprocyanide complexes in preparing test solutions, and
the dissociation of these complexes, like their formation, can be slow.
Incomplete dissociation, as well as unduly high levels of pH or mis-
identification of compounds tested, may have led to some underestimation
of the toxicity of cuprocyanides with high CN/Cu mole ratios.
Doudoroff (1956) and Doudoroff, Leduc, and Schneider (1966) tested
dilutions of relatively concentrated solutions of somewhat uncertain
composition, which were prepared by adding a solution of cupric sulfate,
CuS04, to an NaCN solution until a precipitate (which was not highly
persistent, dissolving eventually) just began to appear. The mole
ratio of initially introduced cyanide to copper was about 3.0. However,
as Doudoroff, Leduc, and Schneider have pointed out, some of the cyanide
initially present must have been lost through oxidation (formation df
cyanogen, which then undergoes hydrolysis, yielding equivalent amounts
of cyanate ion and regenerated free cyanide) as copper(II) was reduced
to copper(I). The final CN/Cu mole ratio thus was believed to have been
some value well above 2.0 but not exceeding 2.5. Even in slightly
alkaline test solutions (pH 7.2-7.9), much free cyanide was present, and
all of the observed toxicity to bluegills of solutions tested by Doudoroff,
Leduc, and Schneider could be attributed to the determined HCN content.
Still, pronounced detoxification of both free cyanide and cupric ion was
shown to result from their reaction when they are combined. Doudoroff
(1956) found the 24-hour median tolerance limit of the mixture for the
fathead minnow in a soft water of nearly neutral reaction to be equivalent
to about 2.2 mg/1 of cyanide (as CN) initially present (total cyanide
introduced), or 1.8 mg/1 of copper. The corresponding values for NaCN
alone and CuS04 alone were 0.25 mg/1 as CN and less than 0.1 mg/1 as Cu,
respectively, under static bioassay conditions. The detoxification of the
combined toxicants was observed both when a fairly concentrated solution
of NaCN (about 940 mg/1 as CN) was mixed with a fairly concentrated solu-
tion of CuS04 (1000 mg/1 as Cu) and when only 0.8 mg of cupric ion per
121
-------
liter was added to an NaCN solution with a cyanide content of only 1.0
mg/1 as CN. However, solutions prepared in the last-mentioned manner
about 1 hour before the introduction of test animals proved somewhat
toxic, killing about half of the fish (55%) in 96 hours, whereas a
comparable dilution of the more concentrated mixture killed no fish in
96 hours, and the 96-hour median tolerance limit of this mixture was
found to be about 1.5 mg/1 as CN initially present, or 1.2 mg/1 as Cu.
Addition to the concentrated mixture of enough CuSO^ or NiSO^ to produce
abundant precipitates (i.e., pronounced turbidity) did not markedly
reduce its toxicity in the moderately soft dilution water.
When diluted with extremely soft, nearly neutral or faintly acid water,
and especially in pure, distilled water (at pH near 6.0), the above-
described, initially concentrated mixture proved decidedly more toxic
than it was in the water that was not quite so soft and that was slightly
alkaline (pH about 7.5) when it was continuously or had been recently
well aerated. Prolonged aeration with compressed air, before the intro-
duction of fish only, or continued also thereafter, of the test solutions
prepared with the very soft water increased their toxicity markedly.
Aging of such a solution in stoppered bottles had a similar but much less
pronounced effect. The behavior and appearance of fish affected or
dying in the solutions prepared with very soft or distilled waters were
suggestive of poisoning with copper, rather than cyanide. Doudoroff
surmised that the toxicity of these solutions probably was due to the
presence of copper ions deriving from dissociation and gradual decomposi-
tion of relatively harmless and stable copper-cyanide complex ions,
although it could not be shown that increased toxicity of the complex
ions in the very soft waters was not a factor. Copper is known to be
extremely toxic to fish in distilled and very soft waters and presumably
was liberated increasingly from cuprocyanide complex ions as free
cyanide (volatile HCN), also deriving from their dissociation, was
removed by prolonged aeration of the neutral solutions. Aeration for 4
122
-------
days (before the fish were introduced) of test solutions containing 0.5
mg/1 of copper combined with cyanide that were prepared with the only
moderately soft water (total alkalinity 17.5 mg/1 as CaCO,; total hard-
*J
ness 25 mg/1 as CaC03; pH during aeration 7.5) did not render these
solutions demonstrably toxic to the minnows. Whether the aeration was
continued or not, the solutions did not prove fatal to the fish within
10 days.
Broderius (1973) studied the toxicity to bluegills of cuprocyanide solu-
tions with CN/Cu mole ratios of 2,0, 2.5, and 3.0, prepared by dissolving
CuCN in very dilute solutions of NaCN in buffered well water, which were
confined in carboys and agitated with magnetic stirrers. Solutions of
varying total cyanide content and having different, widely ranging pH
values (6.51-8.97) were tested, the concentration of molecular HCN in
each one was determined, and the concentrations of the Cu(CN)2~ and
9
complex ions at assumed equilibrium were calculated. It is
noteworthy, however, that equilibria previously had been found not to
have been attained, for reasons not known, in most of the examined solu-
tions with CN/Cu mole ratios of 2.5 or 3.0 even after their prolonged
storage. In solutions like those used in the toxicity tests with a
CN/Cu mole ratio of 2.0, on the other hand, equilibria (indicated by
constancy of HCN levels) were found to be attained fairly rapidly,
within 2 or 3 days; therefore, the results of toxicity tests of the
solutions with this mole ratio, which were used in most of the tests,
are deemed the most meaningful ones. In all of the tested solutions
except two very dilute and acid ones (total CN 0.2 mg/1; pH 6.52-6.57)
with CN/Cu mole ratio of 3.0, in which free cyanide was predominant, the
cyanide group was seen to occur very predominantly in Cu(CN)2~ ions, but
the Cu(CN)3~2 ion apparently can be predominant at equilibrium in other,
similar solutions with a CN/Cu ratio of 3.0 or even less. Concentrations
of the Cu(CN)4~3 ion were believed to have been in all cases negligible,
and in solutions with CN/Cu mole ratio of 2.0, those of the Cu(CN)3
123
-------
ion apparently were and would have remained without exception very low
_2
in comparison with Cu(CN)2~ ion concentrations, though some Cu(CN)j
ion and some remaining CuCN must have been always present. Unreliability
of presently accepted stepwise dissociation constants pertaining to
cuprocyanide complex ions was suggested by Broderius' chemical data, but
the Kp value for the CuCCN^" ion computed by him from equilibrium HCN
levels in his solutions, about 4 x 10, agreed well with previously
published values.
The median survival times of bluegills in the few tested solutions with
CN/Cu mole ratios of 2.5 and 3.0 corresponded closely with those of blue-
gills exposed to NaCN solutions (at pH 7.5) having molecular HCN concen-
trations nearly equal to the determined HCN concentrations in these
cuprocyanide solutions. The acute toxicity of these solutions thus was
clearly attributable to their high molecular HCN content, 0.134-0.269
mg/1 as HCN; HCN then proved lethal to the bluegills in about 11 hours
(median survival time) at a level as low as 0.130 mg/1 in a solution of
NaCN having an initial concentration of 0.15 mg/1 as CN (amount added).
The cuprocyanide solutions had total cyanide concentrations of only 0.2
to 1.0 mg/1 as CN and pH values of 6.5 to 7.5, and the molecular HCN
levels were about one-fifth to nearly four-fifths of the total cyanide
concentrations.
When the CN/Cu mole ratio was 2.0 and the total cyanide concentrations
(5.0 to 50 mg/1 as CN) were more than 20 to 5000 times the molecular HCN
concentrations, the median survival times of the bluegills in the
cuprocyanide solutions were invariably less than those recorded or
estimated for NaCN solutions with the same HCN levels. In solutions with
determined HCN concentrations less than 0.025 mg/1 and total cyanide con-
centrations of 5, 15, 25, and 50 mg/1, the median survival times were, or
averaged, 2520, 632, 354, and 199 minutes, respectively. As the pH of
the solutions of each total cyanide content was reduced and their
124
-------
molecular HCN content was thus increased, their toxicity increased only
slightly, or remained nearly constant, or even declined somewhat (when
the total cyanide concentration was 25 or 50 mg/1) for some undetermined
reason, until a molecular HCN level of about 0.04 to 0.07 mg/1 was
attained. With further decrease of pH and increase of molecular HCN,
the toxicity increased more or less sharply, evidently becoming depend-
ent increasingly on the concentration of HCN and less on the concentra-
tion of cuprocyanide complex ions or total cyanide. These sharp increases
of toxicity occurred when the pH was reduced to levels below a value of
about 7.6-8.0 at all the total cyanide concentrations tested. Even in
the solutions of lower pH and relatively high HCN content, the fish
tended to die sooner than those exposed to the same HCN levels in NaCN
solutions. The undissociated cuprocyanide ions and perhaps also the
copper ions present evidently contributed to the toxicity of these solu-
tions. It should be noted, however, that the total cyanide concentra-
tions in the tested solutions with CN/Cu mole ratio of 2.0 were not very
low; had solutions been tested with much lower total cyanide levels (i.e.,
1.0 mg/1 or less) and pH low enough to cause extensive dissociation of
the Cu(CN)2~ ions and thus to render the solutions acutely toxic, almost
all of their toxicity probably would have been found to be attributable
to their HCN content.
There can be little doubt that the high toxicity of tested solutions
with very low HCN levels (sometimes less than 0.01 mg/1) was due entirely
or almost entirely to toxic action of the cuprocyanide complex ions, the
levels of free cyanide and copper in these solutions having been appar-
ently much too low to have been rapidly effective. The 48-hour median
tolerance limit of the Cu(CN)2~ ion for bluegills in slightly alkaline
solutions in which this complex ion is the only toxicant present in con-
siderable amounts was estimated by Broderius to be about 9 mg/1 as
Cu(CN)2", or 4 mg/1 as CN, at 20° C. By determination of amounts of
copper accumulated in various tissues of bluegills exposed to highly
125
-------
toxic cuprocyanide solutions with very low (presumably negligible)
levels of free cyanide and copper, the Cu(CN)2~ complex ion's ability
to enter the bodies of fish quite rapidly was revealed. The penetration
rate was estimated to be nearly four times that of the silver- cyanide
complex, and this may be the reason why the cuprocyanide ion proved more
toxic than the silver-cyanide complex ion in like toxicity tests. In
the absence of rapidly fatal concentrations of molecular HCN, the
cuprocyanide complex was found to produce signs of intoxication strongly
resembling signs of poisoning with free copper and other heavy metals,
as did also the silver -cyanide complex. Much coagulated mucus was
observed on the gills and body surfaces of the affected fish.
IRON-CYANIDE COMPLEXES
Data in early literature indicating little toxicity to fish of the iron-
cyanide complexes have been contradicted by some more recently published
experimental results but are in agreement with other, even later findings.
Oshima (1931) has reported that 2.5 x 10 M solutions of potassium
f errocyanide , K^FeCCN)^, and potassium ferri cyanide, KgFeCCN)^, killed
young eels, Anguilla japonica, in about 5 to 6 hours, or almost as
rapidly as did a 1 x 10~^ M solution of KCN (0.26 mg/1 free cyanide as
CN). The total cyanide content of a 2.5 x 10"5 M solution of K4Fe(CN)6
or K3Fe(CN)g is only 3.9 mg/1 as CN. On the other hand, Burdick and
Lipschuetz (1950) stated that 4,000 mg/1 of K4Fe(CN)6 (about 1,700 mg/1
as CN), but not 2,000 mg/1, proved fatal to fish (cyprinids) within 24
or 48 hours when the solutions were kept in the laboratory in diffused
light or in the dark. However, only 2 mg/1 (0.85 mg/1 as CN) killed
blacknose dace, Rhinichthys atratulus, creek chubs, Semotilus atromacu-
latus, and silvery minnows, Hybognathus r eg jus, in 1 to 1.5 hours when
the solutions were tested in direct sunlight after their preliminary
exposure to the light for 0.5 to 1.5 hours. Under the same conditions,
2 mg/1 of K3Fe(CN)6 (0.95 mg/1 as CN) killed creek chubs and emerald
126
-------
shiners, Notropis atherinoides, in 13-38 minutes. Bucksteeg (1961) found
that a total cyanide concentration of nearly 300 mg/1 as CN could be
tolerated for 24 hours by rainbow trout in a fresh, solution of K4Fe(CN)6
kept in the dark, but less than 30 mg/1 total CN proved fatal within 24
hours when the solution was exposed to diffused daylight, and only 2
mg/1 caused death within 8 hours when the solution was illuminated with
direct sunlight. The susceptibility of the iron-cyanide complexes to
rapid photolysis (photodecomposition) producing free cyanide must be the
main reason for the large differences of toxicity test results reported
in the literature. The extent of this decomposition in experimental
solutions obviously can vary greatly with the nature or intensity of
illumination of the solutions before and during the tests. Considerable
toxicity of the iron-cyanide complex ions, which slowly dissociate to
some extent even in the dark, has not been demonstrated.
Broderius (1973) found that bluegills lived for more than 48 hours in a
distilled water solution of K^Fe^N)^ that had been kept in the dark in
a closed vessel for a long time (309 days) and had a pH of 7.1, a total
cyanide content of 500 mg/1 as CN, and a determined molecular HCN con-
centration of only 0.067 mg/1, well below the minimum acutely toxic
level. The median survival time of bluegills in a K^Fe^N)^ solution
with the same total cyanide content (500 mg/1 as CN) and the same pH
(7.1), and also thoroughly aged for about 10 months to ensure equilib-
rium, was only 145 minutes. However, the determined molecular HCN con-
tent of this solution was 0.267 mg/1, and the recorded median survival
time of the fish was approximately that which was to be expected at the
determined HCN concentration on the basis of tests of NaCN solutions.
It is quite apparent that the undissociated ferrocyanide ion did not
contribute materially to the lethality of the ferrocyanide solution
(which may have been deficient in dissolved oxygen), although the con-
centration of complexed cyanide was almost 2,000 times that of free
cyanide. Ferrocyanide solutions kept in the dark clearly can become
127
-------
much more toxic to fish, than comparable ferricyanide solutions with the
same total cyanide content, because of greater instability of the former
complex.
Broderius also observed that the molecular HCN concentration in a dilute
K4Fe(CN)6 solution with a total cyanide content of only 5 mg/1 and kept
in the dark increased progressively for about 2 months to nearly 0.06
mg/1. It then declined sharply to a level near 0.01 mg/1 and was nearly
constant thereafter, remaining at that level to the end of the 10-month
experiment. This final concentration was found to be very nearly the
same as the equilibrium concentration in a K^FeCCN)^ solution with a
total cyanide content of 5 mg/1. The pH of both solutions was 6.8 and
the temperature 20° C. Broderius concluded that the decline of the HCN
concentration in the ferrocyanide solution had been due to oxidation of
the iron in the presence of molecular oxygen (02 + 4Fe + 4H+ —>
4Fe+^ + 21^0), the oxidation resulting in complete conversion of the
ferrocyanide to ferricyanide. This explanation was fully supported by
the results of comparative studies of ultraviolet absorption spectra of
the dilute ferrocyanide and ferricyanide solutions stored in the dark
for different periods of time. It was supported also by observed
increases of the pH of the ferrocyanide solution requiring repeated
additions of sulphuric acid for maintenance of a nearly constant pH. The
ultraviolet absorption spectrum of a more concentrated ferrocyanide solu-
tion with a total cyanide content of 500 mg/1 changed only slightly with
time, presumably because of the limited amount of available oxygen, and
the spectra of both dilute and relatively concentrated ferricyanide
solutions remained unchanged.
Broderius noted that the molecular HCN concentrations and the toxicity
of his thoroughly aged ferrocyanide and ferricyanide solutions were not
nearly as great as those predictable through computation of the equilib-
rium HCN levels based on currently accepted values of the cumulative
128
-------
dissociation constants for the complexes. His own "best estimates" of
these constants, derived by computation from the measured equilibrium
levels of molecular HCN, were about 10~47 for the ferrocyariide ion and
52
10~ for the ferricyanide ion at 20° C. These apparent or formal KD
values are smaller by about 11 and 9 orders of magnitude, respectively,
than KD values recently reported in the chemical literature. The value
10~ for the ferrocyanide ion was said to be perhaps slightly too low
an estimate, because of some probable oxidation of the iron in the fairly
concentrated ferrocyanide solution (with 500 mg/1 total CN). The
equilibrium HCN concentration in this solution was used in the computa-
tion without any correction for the probable conversion of some of the
ferrocyanide (a small portion) to ferricyanide. A need for reconcilia-
tion of the earlier published dissociation constants in question with
the observations of Broderius is evident. Some of the possible reasons
for the apparent discrepancy between them already have been noted. The
influence of pH on ferrocyanide or ferricyanide concentrations required
to produce a given level of HCN or toxicity, which can be reasonably
expected to be even more pronounced than the corresponding effect
observed in experiments with the nickelocyanide complex, was not evalu-
ated by Broderius. Such an additional study, and also a study of dis-
sociation rates and equilibrium HCN concentrations in dilute ferrocyanide
solutions devoid of free oxygen, could be quite instructive.
From the information presented above, one can conclude that, were it not
for the phenomenon of photolysis, surface waters receiving moderate
amounts of the ferrocyanide and ferricyanide complexes probably would
never be rendered thereby acutely toxic to fish. Some of the HCN
produced by dissociation of the complexes would be constantly escaping
to the atmosphere or be otherwise lost, and the ferrocyanide complex
would be gradually converted into the more stable ferricyanide complex.
Therefore, and because of the low rate of dissociation of the complexes,
the low equilibrium levels of HCN would never be attained, and the
129
-------
attainment of acutely toxic HCN levels even in waters contaminated with
the ferrocyanide ion would be most improbable. The occurrence of harm-
ful, sublethal levels of HCN deriving from the complexes would still be
possible, for Broderius (1973) has observed HCN concentrations exceeding
0.01 mg/1 even in very recently prepared solutions of both K^FeCCN)^ and
K.jFe(CN)g with a total cyanide concentration of only 5 mg/1. However,
it is mainly because of the susceptibility of the two complexes to
photodecomposition that their disposal in surface waters must be regarded
as potentially harmful or dangerous to fish and other aquatic life. The
computation of equilibrium levels of HCN in solutions not exposed to
light can be of very little or no practical importance.
Burdick and Lipschuetz (1950) determined free cyanide levels in ferro-
cyanide and ferricyanide solutions by a colorimetric method that appar-
ently is sufficiently reliable when only the stable iron-cyanide complexes
are present. They found that a free cyanide concentration of 0.3 mg/1
as CN, which previous tests had shown to produce a 50 percent mortality
of blacknose dace, Rhinichthys atratulus, and creek chubs, Semotilus
atromaculatus, in 5.25 to 7.5 hours, could be attained in solutions of
K4Fe(CN)g and KjFe(CN)^ with a total cyanide content little more than
twice as great by exposing them to direct sunlight. The concentrations
of the two salts required for producing this level of free cyanide in
solutions exposed to direct sunlight in open vessels were reported to
be 1.45 mg/1 K4Fe(CN)6 and 1.34 mg/1 K3Fe(CN)6, i.e., about 0.61-0.64
mg/1 as CN. The authors noted that lower concentrations perhaps would
have been sufficient had the experiments been performed in summer.
Water quality differences were reported to have had a considerable
influence on the liberation of cyanide (i.e., on amounts measured as
free cyanide) in experiments that were alike in other respects.
In reporting concentration values in their paper, Burdick and Lipschuetz
did not distinguish properly between initial ferrocyanide or ferricyanide
130
-------
ion concentrations and initial concentrations by weight of the entire
salt molecules, including the potassium, whose contribution to the weight
of the molecules (unlike that of the cyanide groups) changes with the
valence or oxidation state of the iron. Thus, some values reported or
referred to as ferrocyanide or ferricyanide concentrations (e.g., in
Tables 1 and 2) are elsewhere shown to be concentrations of K4Fe(CN)6
or K3Fe(CN)6 (e.g., in Figure 4). Although comparison of solutions of
the two salts having the same molarity and total cyanide content would
appear to be most appropriate and instructive, after careful examination
of the paper I concluded that all the concentration values given there
(other than free cyanide concentrations) must be concentrations by
weight of the entire molecules, including potassium. Lipschuetz (personal
communication) has informed me that this conclusion is in agreement with
his understanding and recollection. The "cyanide ion" (CN~) concentra-
tions reported are obviously levels of all free cyanide, including HCN.
Burdick and Lipschuetz stated that a part of the phot ode compos it ion of
the iron-cyanide complexes is reversible in the dark or in weaker light,
but no definite evidence of partial reversibility of the reaction was
presented by them. Recombination of free cyanide with other decomposition
products and permanent loss of the cyanide after its liberation in their
tests were not readily distinguishable. Doudoroff (1956) found that very
little if any combination of free cyanide with ferrous ion to form a
complex occurred when only 0.33 mg/1 of free cyanide as CN and 0.13 mg/1
of ferrous ion were introduced into a soft, slightly alkaline water by
the addition of solutions first of NaCN and then of FeS04> A colloidal
precipitate, supposed to have been ferric hydroxide, soon rendered the
water brown, and, after standing for 3 hours, the water remained highly
toxic to fathead minnows, killing all within 24 hours. The ferrocyanide
complex evidently did not form readily also in similar, very dilute solu-
tions prepared by adding first NaCN and then FeS04 to distilled water,
although much complexation of cyanide with iron obviously did occur when
131
-------
much more concentrated solutions of the two salts in distilled water
were combined.
In the British Water Pollution Research Laboratory (Great Britain,
Department of the Environment, 1972), the toxicity to rainbow trout fry
of a saline solution containing 2.0 mg/1 of K4Fe(CN)6 and exposed to
direct sunlight in summer was evaluated by constant-flow bioassay at a
low temperature of 5° C. The reason for performing these tests was that
K4Fe(CN)6 has been added, to prevent caking, to salt used for de-icing
road surfaces in winter. The test solutions were prepared with hard
water to which Nad was added in the amount of 1.54 percent, and they
were irradiated for about 16 hours before the toxicity tests; the
intensity of irradiation was reported to have been about 90,000 lux.
Tests of equally saline solutions of KCN also were performed for compara-
tive purposes. The irradiated ferrocyanide solutions proved somewhat
less toxic than KCN solutions with the same total cyanide content, the
test results suggesting that only about one-half of the cyanide originally
bound in the complex was present as free cyanide, and thus being in good
agreement with the findings of Burdick and Lipschuetz. The "apparent"
72-hour median lethal concentration of K4Fe(CN)g in the irradiated, cold,
saline solution was found to be about 0.17 mg/1. This value is equiva-
lent to 0.07 mg/1 of cyanide as CN, which is approximately the 24-hour
median lethal concentration of free cyanide as CN indicated by the
results of the tests with KCN solutions.
Myers and lezzi (1950) have reported results of some experiments in
which dilute test solutions of K3Fe(CN)6 and K4Fe(CN)6 in a hard water
were irradiated with ultraviolet lamps for not more than 3 hours and
tested for toxicity to bluegills and yellow perch, Perca flavescens.
The 24-hour median tolerance limits of both salts in the irradiated
solutions were found to be usually near 0.5 mg/1 and always less than
0.7 mg/1. There is no indication that the reported values are
132
-------
concentrations of cyanide as CN and not of the entire, double salts.
Inasmuch as 0.5 ing/1 of K3Fe(CN)6 and 0.5 mg/1 of K4Fe(CN)6 are equiva-
lent to 0.24 and 0.21 mg/1 of cyanide as CN, respectively, and the 24-
hour median tolerance limit of KCN was reported by Myers and lezzi to
have been about 0.375 mg/1, or 0.15 mg/1 as CN, liberation of much more
than half of the complexed cyanide in the irradiated solutions is
indicated. If some recombination of the liberated cyanide with other
photodecomposition products occurred after the brief irradiation of the
test solutions, the decomposition could have been almost total, but such
recombination is not indicated by the reported data. When tests were
performed in "ordinary laboratory light", bluegills showed no distress
after exposure for 75 hours to 1000 mg/1 solutions of K3Fe(CN)6 and
K^FeCCN)^, and 50 mg/1 solutions had no effect on yellow perch in 24-
hour tests.
Waters receiving cyanide-bearing wastes usually are not perfectly clear
and very shallow, and often are quite turbid and deep. Because the
penetration of sunlight, and especially that of ultraviolet light, is
limited, photolysis of the iron-cyanide complexes in most of the
receiving waters exposed to sunlight probably is not nearly as rapid as
that observed in aquarium tests. Free cyanide produced by the photolysis
is continuously lost to the atmosphere or otherwise eliminated. There-
fore, concentrations of free cyanide lethal to fish may not be often
attained even in surface waters receiving large amounts of the iron-
cyanide complexes. Fish mortalities apparently attributable to photolysis
of these complexes in streams receiving industrial effluents that were
known to contain the pollutants have been reported, however, by Burdick
and Lipschuetz (1950) and by Myers and lezzi (1950). The source of the
waste involved was not stated by the former authors, who reported a fish
mortality extending over 12 miles (7.5 kilometers) of a stream in the
State of New York but no details of this occurrence. The fish mortalities
reported by Myers and lezzi occurred in the summer of 1949 in Tulpehocken
133
-------
Creek, Lebanon County, Pennsylvania, and were believed to have been
probably due to pollution with wastes from a ferromanganese blast
furnace escaping from a leaking waste-treatment lagoon. No free cyanide
was found in the wastewater and in stream water samples taken at loca-
tions where the death of fish had previously been observed, but "appreci-
able amounts of ferro- and ferricyanide" were said to have been found in
the seepage from the waste-treatment lagoon and in a heavy silt deposit
on the stream bottom. Cyanide was found in the tissues of dead fish in
amounts believed to have been sufficient to have caused death, and local
residents had reported that all the fish mortalities occurred at midday
on bright, sunny days.
It may be helpful to note here that no increase of the toxicity to fat-
head minnows of solutions containing the nickelocyanide and cuprocyanide
complexes upon prolonged exposure of the solutions to direct sunlight
was observed by Doudoroff (1956), and none was to be expected.
Hiatt, Naughton, and Matthews (1953b) reported that the irritant activity
of K^FeCCN)^ on the marine fish Kuhlia sandvicensis was slight at a con-
centration of 1.0 mg/1 and moderate at a concentration of 10 mg/1. Cor-
responding effects of KCN were reported to have been observed at concen-
trations of 0.1 and 1.0 mg/1, respectively. The significance of these
observations is not clear, and their reliability can be reasonably
questioned, in view of the low toxicity of the ferricyanide ion to fresh-
water fishes. It is not evident that experiments with each of the very
numerous chemicals tested were repeated often and carefully enough to
preclude error.
134
-------
SECTION VI
TOXICITY OF OTHER, RELATED COMPOUNDS
NITRILES
The nitriles, or organic cyanides, vary widely in their toxicity to
fish. Acetaldehyde cyanohydrin, CH3CH(OH)CN, also known as lacto-
nitrile, but more appropriately considered as a cyanohydrin because of
its distinctive structure and behavior of toxicological importance,
undergoes rapid decomposition or hydrolysis in aqueous solutions,
yielding CN~ ion and HCN. The toxicity of a solution of this compound
consequently differs little if any from that of a solution of KCN or
NaCN with the same content of the cyanide (CN) group. Daugherty and
Garrett (1951) found the 24-hour median tolerance limit of lactonitrile
for the marine pin perch, Lagodon rhomboides, in unrenewed solutions in
sea water at variable temperatures of 13.7 to 20.4° C to be about 0.22
mg/1 CO.08 mg/1 as CN). Renn (1955) reported 0.51 mg/1 (0.19 mg/1 as
CN) in fresh water to have proved fatal to some bluegills, white
crappie, Pomoxis annularis, and golden shiners, Notemigonus crysoleucas,
in constant-flow tests at 25° C, as well as some static bioassays. The
tests lasted for 24 hours or longer, but deaths of all three species
occurred within 10 hours. Henderson, Pickering and Lemke (1961)
reported both the 24-hour and the 96-hour median tolerance limits at
25° C to have been about 0.90 mg/1 (0.33 mg/1 as CN) for fathead
minnows in both soft and hard waters and for bluegills in soft water,
and 1.37 mg/1 (0.50 mg/1 as CN) for guppies in soft water, when test
solutions were not renewed. When the test solutions were renewed
135
-------
continuously by constant-flow replacement, the 24-hour, 96-hour, 5-day,
and 20-day median tolerance limits for fathead minnows in soft water
were found to be 0.75, 0.71, 0.69, and 0.69 mg/1, respectively (0.27-
0.25 mg/1 as CN).
The 24-hour median tolerance limit of acrylonitrile, CH2CHCN, for the
pin perch was found by Daugherty and Garrett (1951) to be about 25
mg/1 (12 mg/1 as CN), again in tests of unrenewed solutions in sea
water at variable temperatures. Renn (1955) reported that the highest
level that could be tolerated for 24 hours by all of the white crappie
used in his constant-flow tests was between 38 and 68 mg/1 (12 and 25
mg/1 as CN). Bandt (1953) observed that concentrations of 25 to 50
mg/1 (12-25 mg/1 as CN) in standing, aerated solutions eventually
proved lethal to bleak, Alburnus alburnus, and roach, Rutilus rutilus,
in tests at temperatures of 10.5 to 18° C and lasting not longer than
20 days. He concluded that the threshold concentration for prolonged
exposure is in the neighborhood of 20-25 mg/1 (10-12 mg/1 as CN).
When tests were performed without renewal of test solutions, Henderson,
Pickering, and Lemke (1961) found the 24-hour median tolerance limits
at 25° C for fathead minnows in hard water and for fathead minnows,
bluegills, and guppies in soft water to be about 32.7, 34.3, 25.5, and
44.6 mg/1, respectively. The corresponding 96-hour values were found
to be 14.3, 18.1, 11.8, and 33.5 mg/1, respectively (5.8-16.4 mg/1 as
CN). When fathead minnows were tested in continuously renewed solutions
prepared with the soft water, their median tolerance limits for exposure
periods of 1, 2, 4, 10, 15, 20, 25, and 30 days were found to be about
33.5, 14.8, 10.1, 6.9, 5.2, 4.2, 3.5, and 2.6 mg/1, respectively. These
values were obtained by averaging the results of five like experiments
at 25° C in which a total of 50 fish had been exposed to each of seven
different concentrations of the poison. Inasmuch as the logarithms
of the tolerance limits continued to decline almost rectilinearly with
increase of exposure time from 10 to 30 days (also with increase of log
136
-------
time from 10 to 25 days, but seemingly not thereafter), it is evident
that the lethal threshold concentration was not determined and must be
well below 2.6 mg/1 (1.3 mg/1 as CN).
In the British Water Pollution Research Laboratory (Great Britain,
Ministry of Technology, 1970), the median periods of survival of
"trout" (probably rainbow trout) at some unspecified temperature,
presumably in continuously renewed test solutions, were found to be
about 2 and 100 days at acrylonitrile concentrations of 40 and 2.0
parts per million by volume, respectively. Since the density of
acrylonitrile (g/ml) in almost exactly 0.8, these two concentrations
correspond to 32 and 1.6 mg/1 (about 16 and 0.8 mg/1 as CN), respec-
tively, according to my calculations. It was also reported that
exposure of the trout to an acrylonitrile concentration of 60 parts
per million by volume (48 mg/1) caused their delayed death within 10
days after their return to clean, running water. Jackson and Brown
(1970) reported results of very similar if not the same experiments
with rainbow trout, presenting somewhat different data in a graph from
which nearly or exactly the same acrylonitrile concentrations (about 33
and 1.6 mg/1) corresponding to median survival periods of 2 and 100 days
can be derived. In a table in which various toxicity data are summa-
rized, however, the values given by the authors for the 48-hour median
lethal concentration and the concentration that killed 50 percent of
the trout in 100 days (in "hard water") are 70 mg/1 and 2.2 or 2.8 mg/1,
respectively. According to my calculations, these data do not agree
with the data plotted in the graph; yet, I have found no other reason
to believe that they represent results of a different experiment. A
computational error may be involved, or some misunderstanding on my
part. But whatever may be the correct value, it is evident that even
the 100-day median lethal concentration is not nearly equal to the
lethal threshold concentration, which must be considerably lower. The
concentration corresponding to a median survival period of 10 days is
137
-------
about 6.4 mg/1, so the line representing the observed relation between
the logarithms of toxicant concentration and of median survival time
within the range of 2 to 100 days is only slightly curved. Thus, the
median tolerance limit may continue to decrease materially (demon-
strably) with increase of exposure time throughout the normal life span
of the test animal, and it may decline to a value well below 1 mg/1
within that time period. Jackson and Brown stated that deaths of
rainbow trout exposed to an acrylonitrile solution and returned to
clean water occurred some 5 to 10 days later. Temperatures again were
not reported.
At the same laboratory, the toxicity to rainbow trout of malononitrile,
NCCH2CN, which is even more toxic than acrylonitrile and evidently also
an accumulative poison, has been evaluated recently (Great Britain,
Department of the Environment, 1973). Median survival periods of trout
exposed to concentrations of 32, 5, and 0.5 mg/1 were found to be about
5 hours, 32 hours, and 3 days, respectively. The curve relating
survival time to concentration of the poison is unusual, and it suggests
that the lowest concentration tested, 0.5 mg/1 (0-39 mg/1 as CN), is
again much above the lethal threshold concentration, which may be
exceedingly low, perhaps as low as that of free cyanide or even lower.
Henderson, Pickering, and Lemke (1961) tested four other nitriles and
found all of them to be less toxic, and most of them very much less
toxic, than acrylonitrile. Only 96-hour static toxicity bioassays of
these compounds were performed, without renewal of the test solutions,
and the 96-hour median tolerance limits determined were found sometimes
to differ very little or not at all from the corresponding 24-hour
values, perhaps partly because of loss of toxicants from the test
solutions. Each compound was tested with fathead minnows in soft and
hard water and with bluegills and guppies in soft water at a temperature
of 25° C. The four 24-hour median tolerance limits of acetonitrile,
138
-------
CH3CN, ranged from 1050 to 1850 mg/1, and the 96-hour values from 1000
to 1850 mg/1. The 24-hour and 96-hour median tolerance limits of
adiponitrile, NC(CH2)4CN, for fathead minnows differed little; they were
835 and 820 mg/1 in hard water, and 1350 and 1250 mg/1 in soft water.
However, the reported 24-hour values for bluegills and guppies were
1250 and 1200 mg/1, and the corresponding 96-hour values were 720 and
775 mg/1. Oxydipropionitrile, CCH3CH2CN) 20, was the least toxic of the
nitriles tested; its 24-hour median tolerance limits ranged from 4300
to 7350 mg/1, and the 96-hour values from 3600 to 4450 mg/1. Benzo-
nitrile, C6H5CN, proved much more toxic. The 24-hour median tolerance
limits for fathead minnows in hard and soft waters were reported to be
116 and 240 mg/1, respectively, and the corresponding 96-hour values
78 and 135 mg/1, respectively. The 24-hour and 96-hour values.for
each of the other species strangely did not differ at all; they were
78 mg/1 for the bluegill and 400 mg/1 for the guppy.
The principal characteristic of the soft and hard waters used by
Henderson, Pickering and Lemke already have been reported in connection
with the discussion of the influence of water hardness on the lethal
toxicity of free cyanide. The addition only of oxydipropionitrile, of
all the nitriles tested, had a considerable effect on the pH and the
dissolved oxygen content of the waters, the pH of the soft water having
been reduced by almost one pH unit at the highest tested concentration
of this nitrile; the reduction of dissolved oxygen indicated fairly
rapid oxidative degradation of the toxicant. Positive results of
ordinary chemical analysis of the test solutions for cyanide were
obtained only when solutions of lactonitrile and those with high con-
centrations of acetonitrile were tested. The measurable cyanide levels
in the lactonitrile solutions were maximal about 4 hours after the
beginning of the tests and declined almost to nil within 48 hours. The
maxima were nearly equal to the amounts of the-cyanide group introduced.
The measured amounts of cyanide in the solutions of acetonitrile were
139
-------
reported to have been about 0.01 percent of the amounts of the cyanide
group introduced, and not enough to account for the toxicity of the
solutions. The concentrations actually determined in the different
test solutions were not reported. My calculations show that 0.01 per-
cent of the cyanide group present at the reported 96-hour median lethal
concentration of acetonitrile for the bluegill (1850 mg/1) amounts to
about 0.12 mg/1, a concentration not far below the reported 96-hour
median tolerance limit of free cyanide for that species (0.15 mg/1 as
CN). In the case of the fathead minnow, which was reported to be more
sensitive than the bluegill to acetonitrile but less sensitive to free
cyanide, the corresponding difference was much larger. It was reported
that substantial hydrolysis of adiponitrile and oxydipropionitrile
without release of measurable free cyanide was indicated by results of
chemical tests; amounts of ammonia formed were insignificant. No
evidence was observed of similar breakdown of any of the other nitriles
tested.
CYANOGEN CHLORIDE
Cyanogen chloride, CNC1, which is produced upon addition of free
chlorine to cyanide or thiocyanate solutions (e.g., sewage containing
gas plant liquors), appears to be about as toxic to fish as free cyanide,
if not more toxic. Allen, Blezard, and Wheatland (1948) have shown the
lethal threshold concentration for rainbow trout at temperatures of
17-20° C to be below 0.1 mg/1 (0.04 mg/1 as CN). The toxicity tests
were of short duration, apparently lasting not more than 6 hours at
most, and the experimental solutions, prepared with tap water in open
vessels, were not renewed during the tests. The theoretical toxicity
threshold (0.08 mg/1 as CNC1) reported by the authors was derived from
their data by a graphical method that is not current or approved now by
fish toxicologists, but it may approximate the true lethal threshold
concentration.
140
-------
THIOCYANATES
The thiocyanate, CNS~, ion itself is somewhat toxic, but not nearly as
toxic as free cyanide or cyanogen chloride. Brun (1936) observed harm-
ful effects of potassium thiocyanate, KCNS, on Gambusja holbrooki in
10-day tests only at concentrations greater than M/1000 (58 mg/1 as CNS,
or the equivalent of 26 mg/1 of cyanide as CN). Fish exposed to M/500
and M/750 solutions survived for 2 and 6 days, respectively, in these
solutions, but died within 2 and 7 days after their subsequent return
to clean water. A concentration of M/75 proved lethal within 60 hours,
but only a M/25 solution (2320 mg/1 as CMS) killed the fish within 2
hours. The test temperatures are unknown. Schaut (1939) found that
171 mg/1 as KCNS (102 mg/1 as CNS) killed unidentified "minnows" within
9 or 10 days. Herbert (1962) reported that no rainbow trout were killed
in a static bioassay by the highest tested concentration of sodium
thiocyanate, NaCNS, which was 1800 mg/1 as CNS, at 15° C and pH 7.9-8.1,
but the duration of the test was not stated.
Demyanenko (1931) reported that a 1/5000 solution (200 mg/1) of ammonium
thiocyanate, NH4CNS, (153 mg/1 as CNS, or 68 mg/1 as CN) killed bleak,
Albumus alburnus, in about 50 hours, but a concentration half as great
did not prove lethal in 144 hours. A concentration greater by about one-
half (280-300 mg/1 as NH4CNS) was reported by Shelford (1917) to have
been the concentration required to kill orange-spotted sunfish, Lepomis
humilis, in 1 hour. But Oshima (1931) reported that a M/100 (780 mg/1)
solution of NH4CNS (580 mg/1 as CNS, or 260 mg/1 as CN) was tolerated by
young eels, Anguilla japonica, for more than 25 hours, and a concentra-
tion 10 times as great killed the eels only after an exposure of about
3.7 hours. Thumann (1950) observed no effect on two young rainbow trout
after 10 hours at a concentration of 100 mg/1 as NH4CNS, and a single
rainbow trout was not visibly affected at the end of a 135-minute
exposure to 2500 mg/1, but died about 10 days after its return to clean
141
-------
water. A young rainbow trout was killed within 150 minutes by 5000
mg/1, and a young brown trout within 135 minutes by 2500 mg/1.
Wallen, Greer, and Lasater (1957) found the 96-hour median tolerance
limit of NH4CNS for the mosquitofish, Gambusia affinis, in a turbid
water at 16-23° C to be 114 mg/1 (87 mg/1 as CNS, or 39 mg/1 as CN).
However, all the fish died within 144 hours at a concentration of only
56 mg/1 (43 mg/1 as CNS). The 48-hour and 24-hour median tolerance
limits were found to be 420 and 910 mg/1, respectively, and even 1800
mg/1 did not kill most of the fish in 4.5 hours. The pH of the solu-
tions tested by Wallen, Greer, and Lasater was reported to range from
7.4 to 7.9; the pH values of the NH4CNS solutions tested by the other
investigators are unknown. The presence of free, molecular ammonia
(un-ionized base), which is highly toxic to fish, can, of course, cause
or contribute to the acute toxicity of NH^CNS solutions of sufficiently
high pH. However, molecular ammonia is a rapidly fatal poison at con-
centrations not very much greater than the highest concentrations
tolerated for long periods or indefinitely (Wuhrmann and Woker, 1948;
Lloyd, 1961), whereas NH^CNS did not kill fish very rapidly at concen-
trations many times greater than those that proved slowly lethal in the
experiments of Wallen, Greer, and Lasater. Therefore, and in view of
the reported toxicity of KCNS and other pertinent considerations, one
can reasonably conclude that the toxicity of the slowly lethal test
solutions of M^CNS probably was due mostly to the presence of the
thiocyanate ion, which evidently is a slowly acting poison.
Hiatt, Naughton, and Matthews (1953a, 1953b) observed a strong irritant
action of allyl and methyl isothiocyanates and of isobornyl thiocyanate
or thiocyanoacetate on the marine fish Kuhlia sandvicensis. Rapid and
violent reactions leading to dispersal of the schooling fish were
observed at concentrations of these organic chemicals as low as 1.0 mg/1
or less. A moderate (medium) reaction to as little as 0.1 mg/1 of
142
-------
allyl isothiocyanate was reported. In the first account of their study,
Hiatt, Naughton, and Matthews (1953a) reported violent reactions of the
fish also to 20 rag/1 of lauryl thiocyanate and 10 mg/1 of phenyl isothio-
cyanate and of thiocyanic acid 5, 5, 5-trichloro amyl ester. Lower
concentrations produced less violent reactions or were not tested.
Curiously, potassium thiocyanate was reported to have produced a slight
reaction at a concentration of 20 mg/1, but also, in the same publication,
was included (together with ammonium and barium thiocyanates) in a list
of tested chemicals that produced no observable response at that concen-
tration. There are other discrepancies also between results reported in
the two papers cited. No data have been found on the lethality to fish
of any organic thiocyanates.
CYANATES
The cyanate, CNO", ion, which is a product of oxidation of cyanide by
alkaline chlorination, a widely used method of wastewater treatment for
cyanide removal, also appears to be relatively but not entirely harmless
to fish. Washburn (1948) reported the tolerance limit for the creek
chub, Semotilus atromaculatus, of sodium cyanate, NaCNO, to be in the
neighborhood of 75 mg/1 (48.5 mg/1 as CNO, or 30 mg/1 as CN), and
Bucksteeg and Thiele (1957) reported 75 mg/1 as CN to be the lower limit
of concentrations harmful to fish of the potassium salt, KCNO. Enough
information about the experiments performed and the results on which
these values are based has not been included by the authors, however,
in the cited publications. Cyanates may persist in water for a long
time but are subject to hydrolysis yielding ammonium and carbonate ions.
143
-------
SECTION VII
REFERENCES
Abram, F. S. H. 1964. An application of harmonics to fish toxicology.
Int. J. Air Water Pollut. 8_:325-338.
Achard, C., and L. Binet. 1934. Les effets de 1'hyposulfite de soude
sur 1'intoxication par le cyanure de potassium. [The effects of
sodium hyposulfite on intoxication by potassium cyanide.] C. R.
Acad. Sci. 198:222-224.
Alabaster, J. S., J. H. N. Garland, I. C. Hart, and J. F. de L. G.
Solbe". 1972. An approach to the problem of pollution and fish-
eries. Symp. Zool. Soc. London No. 29:87-114.
Alexander, W. B., B. A. Southgate, and R. Bassindale. 1935. Survey of
the River Tees. Part II. The estuary—chemical and biological.
Dept. of Scientific and Industrial Research (Gr. Brit.), Water
Pollution Research, Technical Paper No. 5, H. M. Stationery Office,
London. 171 pp. [Data on temperature effects are presented also
in a summary paper: J. Mar. Biol. Assoc. U. K. 2£: 717-724 (1936).]
Allen, L. A., N. Blezard, and A. B. Wheatland. 1948. Formation of
cyanogen chloride during chlorination of certain liquids. Toxicity
of such liquids to fish. J. Hyg. 46:184-193.
American Public Health Association, et_ al_. 1971. Standard methods for
the examination of water and wastewater. 13th ed. Am. Pub. Health
Assoc., New York. 874 pp. [See pp. 397-398.]
Anderson, P. D. 1974. An approach to the study of multiple toxicity
through the derivation and use of quantal response curves. Ph.D.
thesis, Oregon State University, Corvallis. 80 pp.
Armstrong, C. W. J., and K. C. Fisher. 1940. A comparison of the
effects of the respiratory inhibitors azide and cyanide on the
frequency of the embryonic fish heart. J. Cell. Comp. Physiol.
16:103-112.
144
-------
Bandt, H-J. 1953. Acrylnitril als fischereiliches Abwassergift.
[Acrylonitrile as a waste-water poison affecting fisheries.] Z.
Fisch. Deren Hilfswiss., N.S. 2^:457-461.
Bassindale, R. , B. A. Southgate, and F. T. K. Pentelow. 1933. The
effect of cyanides on the gill colour of fish. J. Mar. Biol.
Assoc. U. K. 18:671-676.
Blaha, J. 1968. Zur Frage der Bestimmung und Toxozitat [sic!) von
freien und komplexen Cyaniden in Wassern. [On the question of
determination and toxicity of free and complex cyanides in waters.]
Vom Wasser 34:175-195.
Bridges, W. R. 1958. Sodium cyanide as a fish poison. U. S. Fish
Wildl. Serv., Spe. Sci. Rep. Fish. No. 253. 11 pp.
Brinley, F. J. 1930. The effect of cyanide on the cardiac rhythm of
embryos of Fundulus heteroclitus. Physiol. Zo81. 3_:283-290.
Brockway, D. L. 1963. Some effects of sub-lethal levels of pentachlor-
phenol and cyanide on the physiology and behavior of a cichlid
fish, Cichlasoma bimaculatum (Linnaeus). M.S. thesis, Oregon State
University, Corvallis. 56 pp.
Broderius, S. J. 1970. Determination of molecular hydrocyanic acid in
water and studies of the chemistry and toxicity to fish of the
nickelocyanide complex. M.S. thesis, Oregon State University,
Corvallis. 93 pp.
Broderius, S. J. 1973. Determination of molecular hydrocyanic acid in
water and studies of the chemistry and toxicity to fish of metal-
cyanide complexes. Ph.D. thesis, Oregon State University,
Corvallis. 287 pp.
Brown, V. M., D. G. Shurben, and D. Shaw. 1970. Studies of water
quality and the absence of fish from some polluted English rivers.
Water Res. £:363-382.
Brun, P. 1936. Sur la toxicite' relative des ions thiocyanique. [On
the relative toxicity of thiocyanate ions.] C. R. Soc. Biol.
121:543-546.
Bucksteeg, W. 1961. Teste zur Beurteilung von Abwa'ssern. [Tests for
the examination of waste waters.] Staedtehygiene 12_(9):180-184.
Bucksteeg, W., and H. Thiele. 1957. Moglichkeiten zur experimentellen
Ermittlung der Einhohnergleichwerte fUr industrielle Abwasser.
[Possibilities of experimental assessment of population equivalents
145
-------
for industrial wastes.] Gas Wasserfach 98_:909-910 (Nasser _
Abwasser pp. 465-466).
Burdick, G. E. 1957. A graphical method for deriving threshold values
of toxicity and the equation of the toxicity curve. N. Y. Fish
Game J. _4:102-108.
Burdick, G. E., H. J. Dean, and E. J. Harris. 1958. Toxicity of
cyanide to brown trout and smallmouth bass. N. Y. Fish Game J.
5_: 133-163.
Burdick, G. E., and M. Lipschuetz. 1950. Toxicity of ferro- and
ferricyanide solutions to fish, and determination of the cause of
mortality. Trans. Am. Fish. Soc. 78[for 1948]:192-202.
Cairns, J., Jr., and A. Scheier. 1958. The effect of periodic low
oxygen upon toxicity of various chemicals to aquatic organisms.
Proc. 12th Ind. Waste Conf., Eng. Bull. Purdue Univ., Eng. Ext.
Ser. No. 94 [Eng. Bull., 42^(3)]: 165-176.
Cairns, J., Jr., and A. Scheier. 1959. The relationship of bluegill
sunfish body size to tolerance for some common chemicals. Proc.
13th Industr. Waste Conf., Eng. Bull. Purdue Univ., Eng. Ext. Ser.
No. 96[Eng. Bull. 43_ (3)]: 243-252.
Cairns, J., Jr., and A. Scheier. 1963. Environmental effects upon
cyanide toxicity to fish. Not. Nat. Acad. Nat. Sci. Philadelphia,
No. 361. 11 pp.
Cairns, J., Jr., and A. Scheier. 1968. A comparison of the toxicity
of some common industrial waste components tested individually and
combined. Prog. Fish-Cult. 30:3-8.
Cairns, J., Jr., A. Scheier, and J. J. Loos. 1965. A comparison of the
sensitivity to certain chemicals of adult zebra danios Brachydanio
rerio (Hamilton-Buchanan) and zebra danio eggs with that of adult
bluegill sunfish Lepomis machrochirus Raf. Not. Nat. Acad. Nat.
Sci. Philadelphia, No. 381. 9 pp.
Carter, L. 1962. Bioassay of trade wastes. Nature (London) 196:1304.
Chan, K. K-S. 1971. Some effects of chronic cyanide poisoning on
osmoregulation of rainbow trout. M.S. thesis, Sir George Williams
University, Montreal, Canada. [Abstract only seen.]
Chen, C. W., and R. E. Selleck. 1969. A kinetic model of fish toxicity
threshold. J. Water Pollut. Control Fed. 41(8, Pt.2):R294-R308.
146
-------
Christensen, J. J., R. M. Izatt, J. D. Hale, R. T. Pack, and G. D. Watt.
1963. Thermodynamics of metal cyanide coSrdination. II. AG°,
AH°, and AS0 values for tetracyanoniccolateCH) ion formation in
aqueous solution at 25°. Inorg. Chem. 2^:337-339.
Claeys, R. R., and H. Freund. 1968. Gas chromatographic separation of
HCN on Porapak Q: Analysis of trace aqueous solutions. Environ.
Sci. Technol. 2^:458-460.
Costa, H. H. 1965. Responses of freshwater animals to sodium cyanide
solutions. 1. Fish. Ceylon J. Sci. Biol. Sci. _5(2):41-87.
Costa, H. H. 1966. The effect of exercise on the survival of Phoxinus
phoxinus L. in sodium cyanide solutions. Hydrobiologia 28(2) : 241-
251. —
Daugherty, F. M., and T. J. Garrett. 1951. Toxicity levels of
hydrocyanic acid and some industrial by-products. Tex. J. Sci.
_3:391-396.
Demyanenko, V. N. 1931. Poisoning of fish by waste waters from
chemical factories and the "fish test" [in Russian]. Gig.
Epidemiol. 10(6/7):13-20.
Dorier, A. 1952. Toxicite pour les poissons de cyanure de sodium
industriel. [Toxicity to fish of industrial sodium cyanide.]
Trav. Lab. Hydrobiol. Pise. Grenoble 43_-44[for 1951-52]: 103-106.
Doudoroff, P. 1956. Some experiments on the toxicity of complex
cyanides to fish. Sewage Ind. Wastes 28:1020-1040.
Doudoroff, P. 1965. Formal Discussion (of paper by S. Ishio, "Behavior
of fish exposed to toxic substances"). Pages 33-35 iii 0. Jaag, ed.
Advances in water pollution research (Proc. Int. Conf. Int. Assoc.
Water Pollut. Res.), 1964, Vol. 1. Permagon Press, Inc., New York.
Doudoroff, P., and M. Katz. 1950. Critical review of literature on
the toxicity of industrial wastes and their components to fish.
I. Alkalies, acids, and inorganic gases. Sewage Ind. Wastes
2^:1432-1458.
Doudoroff, P., G. Leduc, and C. R. Schneider. 1966. Acute toxicity to
fish of solutions containing complex metal cyanides, in relation
to concentrations of molecular hydrocyanic acid. Trans. Am. Fish.
Soc. 95:6-22.
Downing, K. M. 1954. The influence of dissolved oxygen on the toxicity
of potassium cyanide to rainbow trout. J. Exp. Biol. 31_: 161-164.
147
-------
Ellis, M. M. 1937. Detection and measurement of stream pollution.
U. S. Bur. Fish. Bull. No. 22. Bull. Bur. Fish. 48^:365-437.
Fisher, K. C., and R. Ohnell. 1940. The steady state frequency of the
embryonic fish heart at different concentrations of cyanide. J.
Cell. Comp. Physiol. 16:1-15.
Freund, H., and C. R. Schneider. 1959. Determination of the cumulative
dissociation constant of tetracyanonickelate(II) ion. J. Am. Chem.
Soc. 81^:4780-4783.
Gillar, J. 1962. The effect of cyanide on some aquatic animals [in
Czech, with Russian and German summaries]. Sb. Vys. Sk. Chem.-
Technol. Praze, Technol. Vody 6_(Pt. l):435-457.
Grau, P., and J. Hrubec. 1965. Reasons and course of the disastrous
pollution of the Jihlava River with cyanide [in Czech, with English
and Russian summaries separately provided]. Vodni Hospod.
15.C1): 19-20.
Great Britain, Department of Scientific and Industrial Research. 1953.
Water pollution research 1952. H. M. Stationery Office, London.
[See pp. 42-45.]
Great Britain, Department of Scientific and Industrial Research. 1956.
Water pollution research 1955. H. M. Stationery Office, London.
[See pp. 37-38.]
Great Britain, Department of the Environment. 1972. Water pollution
research 1971. H. M. Stationery Office, London. [See pp. 38-40.]
Great Britain, Department of the Environment. 1973. Water pollution
research 1972. H. M. Stationery Office, London. [See pp. 39-41.]
Great Britain, Ministry of Technology. 1967. Water pollution research
1966. H. M. Stationery Office, London. [See pp. 152-154.]
Great Britain, Ministry of Technology. 1968. Water pollution research
1967. H. M. Stationery Office, London. [See pp. 60-65.]
Great Britain, Ministry of Technology. 1969. Water pollution research
1968. H. M. Stationery Office, London. [See pp. 61 and 65.]
Great Britain, Ministry of Technology. 1970. Water pollution research
1969. H. M. Stationery Office, London. [See pp. 63-64.]
Henderson, C., Q. H. Pickering, and A. E. Lemke. 1961. The effects of
some organic cyanides (nitriles) on fish. Proc. 15th Ind. Waste
148
-------
" Bull- Purdue Univ" Engl Ext- Ser- N°- 106 EEng- Bul1-
Herbert, D. W. M. 1962. The toxicity to rainbow trout of spent still
liquors from the distillation of coal. Ann. Appl. Biol. 5£:755-
/ / / *
Herbert, D. W. M. , and K. M. Downing. 1955. A further study of the
toxicity of potassium cyanide to rainbow trout (Salmo gairdneri
Richardson). Ann. Appl. Biol. 43:237-242.
Herbert, D. W. M. , K. M. Downing, and J. C. Merkens. 1955. Studies on
the survival of fish in poisonous solutions. Int. Ver. Theor.
Angew. Limnol., Verh. 12:789-794.
Hiatt, R. W. , J. J. Naughton, and D. C. Matthews. 1953a. Effects of
chemicals on schooling fish, Kuhlia sandvicensis. Biol. Bull.
(Woods Hole, Mass.) 104_:28-44T
Hiatt, R. W. , J. J. Naughton, and D. C. Matthews. 1953b. Relation of
chemical structure to irritant responses in marine fish. Nature
(London) 172:904-905.
^
Hubault, E. 1955a. Etude sur la progressivit^ de la pollution de la
Meurthe en fonction du debit et sur les seuils de nocivit£ de
divers composes chimiques vis-a-vis du poisson. [Study of the
progress of pollution of the Meurthe in relation to the flow and
of the toxicity thresholds of various chemical compounds for fish.]
Eau 42^(11): 2 71 -2 76.
Hubault, E. 1955b. Les seuils de nocivite" de divers composes chimiques
vis-a-vis du poisson. [The toxicity thresholds of various chemical
compounds for fish.] C. R. 27th Congr. Int. Chim. Ind. (Brussels)
1954 !_; Ind. Chim. Beige 2£ (Spec. No.) :352-356.
Ishida, J. 1947. Mechanism of opercular movement by the embryo of
killifish [in Japanese]. Dobutsugaku Zasshi 57_(l/2) : 14-15. Biol.
Abs. (1950) 24_:635, Abs. No. 6446. [Indirectly cited.]
Ishio, S. 1965. Behavior of fish exposed to toxic substances. Pages
19-33 in 0. Jaag, ed. Advances in water pollution research (Proc.
Int. Conf. Int. Assoc. Water Pollut. Res.), 1964, Vol. 1.
Permagon Press, Inc., New York.
Jackson, S., and V. M. Brown. 1970. Effect of toxic wastes on treat-
ment processes and watercourses. Water Pollut. Control 69_:292-
303.
149
-------
Jones, J. R. E. 1947. The oxygen consumption of Gasterosteus
aculeatus L. in toxic solutions. J. Exp. Biol. 23:298-311.
Kariya, T., R. Akiba, S. Suzuki, and T. Tsuda. 1967. Studies on the
post-mortem identification of the pollutant in the fish killed by
water pollution. IV. Detection of cyanide in fish [in Japanese,
with English abstract, table headings, and figure legends]. Bull.
Jpn. Soc. Sci. Fish. (Tokyo) 33_(4) :311-314.
Karsten, A. 1934. Investigation of the effect of cyanide on Black
Hills trout. Black Hills Eng. £2:145-174.
Leclerc, E., and F. Devlaminck. 1950. Etude toxicologique de quelques
substances ge'ne'ralement presents dans les effluents d'usines & gaz.
[Toxicological studies of some substances usually present in
effluents from gas works.] Cent. Beige Etude Doc. Eaux, Bull.
Mens. No. 8 [1950/11]=486-493.
Leduc, G. 1966. Some physiological and biochemical responses of fish
to chronic poisoning by cyanide. Ph.D. thesis, Oregon State
University, Corvallis. 146 pp.
Leschber, R. 1969. Die Beurteilung der Toxizitat cyanidhaltiger
AbwSsser. Zur Frage der Beziehung zwischen der Toxizitat und der
analytischen Bestimmung von Cyaniden durch geeignete Verfahren.
[The estimation of the toxicity of cyanide-containing wastewaters.
On the question of the relation between the toxicity and the
analytical determination of cyanides by appropriate methods.]
Galvanotechnik 60:368-374.
Lewis, W. M., and R. M. Tarrant, Jr. 1960. Sodium cyanide in fish
management and culture. Prog. Fish-Cult. 22:177-180.
Lipschuetz, M., and A. L. Cooper. 1955. Comparative toxicities of
potassium cyanide and potassium cuprocyanide to the western black-
nosed dace CRhinichthys atratulus meleagris). N. Y. Fish Game J.
£:194-204.
Llyod, R. 1961. The toxicity of ammonia to rainbow trout. Water Waste
Treatment J. 8:278-279.
Lloyd, R., and D. H. M. Jordan. 1963. Predicted and observed toxicities
of several sewage effluents to rainbow trout. Inst. Sewage Purif.,
J. Proc. 1963 (Pt.2):167-173.
Lloyd, R., and D. H. M. Jordan. 1964. Predicted and observed toxicities
of several sewage effluents to rainbow trout: A further study.
Inst. Sewage Purif., J. Proc. 1964 (Pt. 2):183-186.
150
-------
Malacca, I. 1966. Contributions to knowledge of the toxic effects of
cyanides, ammonia, mercury, and arsenic on some species of fish
and on Daphnia [in Roumanian, with English, German, and Russian
summaries]. Stud. Prot. Epurarea Apelor 7J2') : 751-792.
Malacea, I. 1968. Untersuchungen iiber die Gewohnung der Fische an
hohe Konzentrationen toxischer Substanzen. [Studies on the accli-
matization of fishes to high concentrations of toxic substances.]
Arch. Hydrobiol. 65:74-75.
Michigan Department of Conservation, Institute for Fisheries Research.
1933. The toxicity to fish life of cyanide solutions, and of the
products derived from certain chemical treatments of cyanide
solutions. Rep. No. 207, Inst. Fish. Res., University of Michigan,
Ann Arbor. Typescript or mimeo. (photostatic copy seen). 14 pp.
Milne, D. 1950. Disposal of cyanides by complexation. Sewage Ind.
Wastes 22:1192-1199.
Montgomery, H. A. C. , D. K. Gardiner, and J. G. G. Gregory. 1969.
Determination of free hydrogen cyanide in river water by a solvent-
extraction method. Analyst (London) 94^:284-291.
Moore, S. L. , and S. R. Kin. 1968. Cyanide pollution and emergency
duty, train wreck, Dunreith, Indiana. Proc. 23rd Industr. Waste
Conf., Eng. Bull. Purdue Univ., Eng. Ext. Ser. No. 132 (Pt. 1)
[Eng. Bull. 53^(2)]:583-600.
Myers, C. S., and T. lezzi. 1950. A report of experiments on the
effect of certain toxicants on fish life with especial reference to
the effect of photodecomposition. Mimeo. Rep., Ind. Wastes Div.,
Bur. Sanit. Eng., Pennsylvania Department of Health, Harrisburg,
Pa. 10 pp.
Negilski, D. S. 1973. Individual and combined effects of cyanide,
pentachlorphenol and zinc on juvenile chinook salmon and inverte-
brates in model stream communities. M.S. thesis, Oregon State
University, Corvallis. 80 pp.
Nehring, D. 1964. Die Schadwirkung von Kupfersulfat, Zinksulfat,
Kaliumzyanid, Ammoniak und Phenol gegeniiber Karpfen (Cyprinnus
carpio) vom Wasser her und nach peroraler Applikation~[The
toxicity to carp (Cyprinus carpio) of copper sulfate, zinc sulfate,
potassium cyanide, ammonia, and phenol from water and after oral
administration.] Z. Fisch. Deren Hilfswiss., N.S. 12^:717-724.
Neil, J. H. 1957. Some effects of potassium cyanide on speckled trout
CSalvelinus fontinalis). Pages 74-96 in_ Papers presented at fourth
151
-------
Ontario industrial waste conference. Water and Pollution Advisory
Committee, Ontario Water Resources Commission, Toronto, Canada.
Nelson, K. H., and I. Lysyj. 1971. Analysis of water for molecular
hydrogen cyanide. J. Water Pollut. Control Fed. 45:799-805.
Oshima, S. 1931. On the toxic action of dissolved salts and electro-
lytes upon young eels (Anguilla japonica) [in Japanese]. J. Imp.
Fish. Exp. Sta. (Tokyo) No. 2:139-193.
Patrick, R., J. Cairns, Jr., and A. Scheier. 1968. The relative
sensitivity of diatoms, snails, and fish to twenty common constitu-
ents of industrial wastes. Prog. Fish-Cult. 3Jh 137-140.
Phillips, F. S. 1940. Oxygen consumption and its inhibition in the
development of Fundulus and various pelagic fish eggs. Biol. Bull.
(Woods Hole, Mass.) 78_:256-274.
Powers, E. B. 1917. The goldfish (Carassius carassius) as a test
animal in the study of toxicity. 111. Biol. Monogr. 4_(4) :123-193
[or pp. 3-73 of No. 4 only].
Renn, C. E. 1955. Biological properties and behaviors of cyanogenic
wastes. Sewage Ind. Wastes 27:297-308.
Schaut, G. G. 1939. Fish catastrophies during droughts. J. Am. Water
Works Assoc. 51_: 771-821.
Schneider, C. R., and H. Freund. 1962. Determination of low level
hydrocyanic acid in solution using gas-liquid chromatography.
Anal. Chem. 54:69-74.
Seeberger, X. 1921. Toxische Wirkung von Brennereiruckstanden auf
Fische. [Toxic action of distillery wastes on fish.] Verh.
Schweiz. Naturforsch. Ges. 102_(Pt. 2):193-195.
Seth, A. K., S. K. Shrivastava, M. G. George, and J. K. Bewtra. 1967.
Monitoring of certain toxic constituents in water supplies.
Environ. Health 9_:34-38.
Shelford, V. E. 1917. An experimental study of the effects of gas
wastes upon fishes, with especial reference to stream pollution.
Bull. 111. State Lab. Nat. Hist. 11_:381-412 + figure and charts
with captions (12 pp).
Silaichuk, E. V. 1969. Effect of waste waters containing cyanides and
hexavalent chromium on the survival of certain aquatic organisms
[in Russian]. Gig. Sanit. 34J12) :78-79.
152
-------
Southgate, B. A. 1932. The toxicity of mixtures of poisons. Q. J
Pharm. Pharmacol. 5:639-648.
Southgate, B. A., F. T. K. Pentelow, and R. Bassindale. 1932. An
investigation into the causes of death of salmon and sea trout
smolts in the estuary of the River Tees. Biochem. J. 26^273-284.
Southgate, B. A., F. T. K. Pentelow, and R. Bassindale. 1933. The
toxicity to trout of potassium cyanide and p-cresol in water
containing different concentrations of dissolved oxygen. Biochem.
J. 27^:983-985.
Sprague, J. B. 1970. Measurement of pollutant toxicity to fish. II.
Utilizing and applying bioassay results. Water Res. £:3-32.
Stumm, W., H. Woker, and H. U. Fischer. 1954. Die Entgiftung von
zyanidhaltigen Abwassern durch Oxydation mit Hypochlorit.
[Detoxification of waste waters containing cyanide by oxidation
with hypochlorite.] Schweiz. Z. Hydrol. 16:1-21.
Summerfelt, R. C., and W. M. Lewis. 1967. Repulsion of green sunfish
by certain chemicals. J. Water Pollut. Control Fed. 59:2030-2038.
Sunmer, F. B. , and P. Doudoroff. 1938. Some experiments on temperature
acclimatization and respiratory metabolism in fishes. Biol. Bull.
CWoods Hole, Mass.) 74_: 403-429.
Sumner, F. B., and M. C. Sargent. 1940. Some observations on the
physiology of warm spring fishes. Ecology 21:45-54.
Sumner, F. B., and N. A. Wells. 1935. Some relations between respira-
tory metabolism in fishes and susceptibility to certain anesthetics
and lethal agents. Biol. Bull. (Woods Hole, Mass.) 6£:368-378.
Symons, G. E. , and R. W. Simpson. 1939. Report on fish destruction in
the Niagara River. Trans. Am. Fish. Soc. 68[for 1938]:246-255.
Thumann, M.-E. 1950. liber die Wirkung von AmmoniumsaltzlSsungen auf
Regenbogen- und Bachforellen und einige Fischnahrtiere. [On the
action of ammonium salt solutions on rainbow and brown trout and
on some fish-food organisms.] Abh. Fisch. Deren Hilfswiss. 1950,
Pt. 2:327-348.
Turnbull, H., J. G. DeMann, and R. F. Weston. 1954. Toxicity of
various refinery materials to fresh-water fish. Ind. Eng. Chem.
46:324-333.
153
-------
Wallen, I. E., W. C. Greer, and R. Lasater. 1957. Toxicity to Gambusia
affinis of certain pure chemicals in turbid waters. Sewage Ind.
Wastes 29.: 695-711.
Warren, C. E. (with P. Doudoroff). 1971. Biology and water pollution
control. W. B. Saunders Co., Philadelphia. 434 pp.
Washburn, G. N. 1948. The toxicity to warm-water fishes of certain
cyanide plating and carburizing salts before and after treatment
by the alkali-chlorination method. Sewage Works J. 20:1074-1083.
Weiss, B., H. A. Abramson, and M. 0. Baron. 1958. Lysergic acid
diethylamide (LSD-25). XXV. Effect of potassium cyanide and
other oxidase and respiratory inhibitors on the Siamese fighting
fish. Arch. Neurol. Psychiatry 80:345-550.
Wells, M. M. 1916. Starvation and the resistance of fishes to lack of
oxygen and to KCN. Biol. Bull. (Woods Hole, Mass.) 31:441-452.
Woker, H., and K. Wuhrmann. 1950. BeitrSge zur Toxicologie der Fische.
VI. Die EmpfindlichJceit verschiedener Fischarten gegeniiber
Ammoniak, BlausSure und Phenol. [Contributions to fish toxicology.
VI. The sensitivity of different species of fish to ammonia,
hydrocyanic acid, and phenol.] Rev. Suisse Zool. 57(3):548-555.
Worley, F. P., and V. R. Browne. 1917. The hydrolysis of sodium
cyanide. J. Chem. Soc. (Trans.) 111:1057-1062.
Wufirmarai, K. 1952. Sur quelques principes de la toxicologie du poisson.
(Concerning some principles of the toxicology of fishes.) Cent.
Beige Etude Doc. Eaux, Bull. Mens. No. 15 [1952/1]:49-60.
Wuhrmann, K., and H. Woker. 1948. Beitrage zur Toxicologie der Fische.
II. Experimentelle Untersuchungen iiber die Ammoniak- und
BliausSurevergiftung. [Contributions to fish toxicology. II.
Experimental investigations of ammonia and hydrocyanic acid
poisoning.] Schweiz. Z. Hydrol. 11:210-244.
Wuhrmann, K.A and H. Woker. 1953. Beitra'ge zur toxicologie der Fische.
VIII". IJber diet Glf twirkungen von Ammoniak- und Zyanidio'sungen mit
verschiedenex. Saaierstoffspannung und Temperatur auf Fische.
[ContributLiJon^s to: fish toxicology. VIII. Concerning the toxicity
to fishes oft ammonia and cyanide solutions with varying oxygen
tension and temperature.] Schweiz. Z. Hydrol. 15:235-260.
Wuhrmann, K., and H. Woker. 1955. Influence of temperature and oxygen
tension on the toxicity of poisons to fish. Int. Ver. Theor.
Angew. Limnol., Verh. 12:795-801.
154
-------
TECHNICAL REPORT DATA
(Please read Instructions on the reverse before completing)
1. REPORT NO.
EPA-600/3-76-058
3. RECIPIENT'S ACCESSIOr*NO.
4. TITLE AND SUBTITLE
TOXICITY TO FISH OF CYANIDES AND RELATED COMPOUNDS
A REVIEW
5. REPORT DATE
April 1976 Tissuing Date)
6. PERFORMING ORGANIZATION CODE
N/A
7. AUTHOR(S)
Peter Doudoroff
8. PERFORMING ORGANIZATION REPORT NO.
N/A
|9. PERFORMING ORGANIZATION NAME AND ADDRESS
Department of Fisheries and Wildlife
Oregon State University
Corvallis, Oregon 97331
10. PROGRAM ELEMENT NO.
1BA608
11. CONTRACT/GRANT NO.
Grant R-802459
12. SPONSORING AGENCY NAME AND ADDRESS
U.S. Environmental Protection Agency
Office of Research and Development
Environmental Research Laboratory
Ouluth, Minnespta 55804
13. TYPE OF REPORT AND PERIOD COVERED
Final
14. SPONSORING AGENCY CODE
EPA-ORD
15. SUPPLEMENTARY NOTES
16. ABSTRACT
The world literature on the toxicity to fish of simple and complex cyanides, nitriles,
cyanogen chloride, thiocyanates, and cyanates is reviewed critically and interpre-
tively. Differently determined limits of toxicant concentrations tolerated by various
fishes are compared, and their variation with exposure time, the pH, temperature, and
dissolved oxygen and mineral content of the water, body size, age, acclimation, etc.,
is examined. Interactions of free cyanide with other toxic water pollutants also are
considered. Available data on effects of sublethal levels of free cyanide on growth,
food consumption and utilization, swimming ability, behavior, etc., and observations
on avoidance reactions of fish to the toxicant are summarized and their ecological
significance is discussed. After a brief introduction to the chemistry of complex
metallocyanides and their behavior in dilute solutions, the acute toxicity of the
solutions is thoroughly considered and related to concentrations of their identifiable
components. The dominant role of molecular hydrocyanic acid produced by dissociation
or photolysis of the metallocyanide complexes as a lethal agent responsible for the
toxicity of most of the toxic solutions tested is given particular attention; the
relative toxicity of complex metallocyanide ions also is considered. Some conclusions
regarding acceptable concentrations of free cyanide in receiving waters are presented.
17.
KEY WORDS AND DOCUMENT ANALYSIS
DESCRIPTORS
b. IDENTIFIERS/OPEN ENDED TERMS
COS AT I Field/Group
Nitriles
Thiocyanates
PH
Temperature
Growth
Food consumption
Behavior
Reviews
Toxicity
Cyanides
Interactions
Hydrogen cyanide
Fish toxicity
Cyanogen chloride
Exposure time
Swimming ability
Mettalocyanide
Concentrations acceptable
06F
06S
06C
3. DISTRIBUTION STATEMENT
Release to Public
IMMMVMMM_^HMHHM_
EPA Form 2220-1 (9-73)
19. SECURITY CLASS (ThisReport)
Unclassified
21. NO. OF PAGES
161
20. SECURITY CLASS (Thispage)
22. PRICE
Unclassified
155
^^mmma^mmm^mm i i i
6USGPO: 1976 — 657-695/5403 Region 5-11
------- |