Algal Assay Procedure
Bottle Test
NATIONAL EUTROPHICATION RESEARCH PROGRAM
ENVIRONMENTAL PROTECTION AGENCY
-------
ALGAL ASSAY PROCEDURE
BOTTLE TEST
National Eutrophication Research Program
Environmental Protection Agency
-------
PLEASE NOTE
Enclosed U a copy oi the ALGAL ASSAY PROCEVURE: BOTTLE TEST.
We expect to update the, pnoceduAe fiAorn time to .time.. To
facilitate this, ooe would appreciate Ktc.elvx.ng any comments
ok suggestions you may have fi.egan.dU.ng Its use and application.
flease send these to-
Wl, Thomas E. Moloney, Acting Chief
National Eu&iophlcatlon Research ?n.ogKam
National Environmental Research Centex.
ZOO SW 3Sth Stsieet
-------
TABLE OF CONTENTS
Page
FOREWORD i
PARTICIPATING LABORATORIES ii
INTRODUCTION 1
PRINCIPLE 3
PLANNING AND EVALUATION OF ALGAL ASSAYS 4
Selection of Test Waters 5
Determination of Limiting Nutrients 5
Evaluation of Materials 6
Assessment of Waste Treatment Processes 7
APPARATUS 7
Sampling and Sample Preparation 7
Culturing and Incubation 7
General 9
Optional 9
SAMPLE COLLECTION, TRANSPORT, PREPARATION AND STORAGE 9
Collection 9
Transport Conditions 9
Preparation 9
Storage 11
SYNTHETIC ALGAL NUTRIENT MEDIUM 11
Final Concentration of Nutrients 11
Stock Solutions 12
Preparation of Medium I2
INOCULUM 13
Test Algae 13
Source of Test Algae 13
Maintenance of Stock Cultures 13
Preparation of Inoculum 14
-------
TABLE OF CONTENTS (Cont'd)
Pa^e
TEST CONDITIONS 15
Temperature 15
I llutni nation 15
PROCEDURE 15
Preparation of Glassware 15
pH Control 16
Growth Parameters 16
Maximum Specific Growth Rate 17
Maximum Standing Crop 18
Biomass Monitoring 19
SPIKES 20
Auxiliary Spikes 21
DATA ANALYSIS AND INTERPRETATION 21
Introduction 21
Reference Curves 21
Maximum Specific Growth Rate 22
Maximum Standing Crop 22
Confidence Intervals 23
Rejection of Outliers 23
Evaluation of Assay Results 23
APPENDICES 27
Determination of Limiting Nutrients 27
Evaluation of Materials 30
Example of Method to Assess Waste Treatment Processes ... 31
Sample Preparation 35
Taxonomy, Morphology, and Reproduction of Algal Test Species . 41
Test Algae 41
Illumination 46
Biomass Monitoring 51
Dry Weight 51
Direct Microscopic Counting 54
Absorbance 54
Chlorophyll 56
Data Analysis and Interpretation 62
Reference Curves 62
Maximum Specific Growth Rate 62
Maximum Standing Crop 62
Confidence Intervals 62
Rejection of Outliers 75
-------
FOREWORD
This document is the product of more then two years of intensive
research at governmental, industrial, and university laboratories
to develop a reliable and reproducible algal assay. The "Bottle
Test" of the Provisional Algal Assay Procedure (PAAP), which was
published by the Joint Industry/Government Task Force on Eutrophication
in February, 1969, served as the starting point for this effort. In
improving and evaluating the test, each laboratory, of the participating
group of nine, followed the same research plan using algal test species
from a common source. This course of action identified elements of
the test that were faulty or difficult or questionable. As such
specific problems were recognized, selected laboratories were assigned
the task of probing them and developing sound adjustments. To do
this they followed research plans developed jointly by all laboratories.
As a result of this massive effort, the Algal Assay Procedure:
Bottle Test has been refined sufficiently to be offered now for wider
use in connection with eutrophication and other algal production problems.
This point in progress has been attained only through the intense interest
and continuing energies of the participating laboratories. These
laboratories and the personnel who worked on this project are shown on
the following page.
Coordination of this program has been the responsibility of
Mr. Thomas E. Maloney of the National Eutrophication Research Program,
Pacific Northwest Water Laboratory, Corvallis, Oregon. To him goes
much credit for the effective way in which the program has moved
continuously and effectively toward this goal.
Finally, the way is now open to move on with the next step-
that of learning to use this newly improved procedure as an aid in
solving practical problems. This publication is one step in stimulating
such action.
A. F. Bartsch, Director
Pacific Northwest Water Laboratory
-------
PARTICIPATING LABORATORIES AND PERSONNEL
UNIVERSITY OF CALIFORNIA,
BERKELEY
Erman A. Pearson
Donald B. Porcella
Daan F. Toerien
J. Radimsky
C. H. Huang
Peter Grau
LEVER BROTHERS COMPANY,
EDGEWATER, N. J.
Warren G. Yeisley
Eugene Mones
Irving M. Williams, Jr.
FMC,
PRINCETON, N. J.
Eckhardt G. Lindemann
Gert P. Volpp
Paul F. Derr
William W. Smith
Jane C. Warner
MONSANTO COMPANY,
ST. LOUIS, MISSOURI
Dee Mitchell
Richard Kimerle
PROCTER AND GAMBLE COMPANY,
CINCINNATI, OHIO
James R. Duthie
James E. Thompson
A. G. Payne
J. W. Williams
UNIVERSITY OF WISCONSIN,
MADISON
Gerard A. Rohlich
Paul D. Uttormark
Robert M. Gerhold
George P. Fitzgerald
Cynthia R. Nadler
UNIVERSITY OF CALIFORNIA,
IRVINE
Jan W. Scherfig
Peter S. Dixon
Carol Justice
Steven N. Murray
UNIVERSITY OF NORTH CAROLINA,
CHAPEL HILL
SCHOOL OF PUBLIC HEALTH
Department of Environmental Sciences
and Engineering
Charles M. Weiss
J. Benjamin SI ess
James R. Haviland
Frank G. Wilkes
Gail H. Marshall
Department of Biostatisties
Ronald W. Helms
NATIONAL EUTROPHICATION RESEARCH
PROGRAM, CORVALLIS, OREGON
Thomas E. Maloney
William E. Miller
Tamotsu Shiroyama
Stephanie Vreeland
-------
ALGAL ASSAY PROCEDURE: BOTTLE TEST
1. Introduction
With the surge of interest in the growing problem of eutrophication,
the Joint Industry/Government Task Force on Eutrophication recognized
that acceptable standardized algal growth tests must be developed as a
tool in controlling eutrophication. While many investigators have
improvised algal assays to meet their specific needs, these assays offer
no basis for comparison of acceptably reproducible results between
laboratories or on samples obtained from different geographic areas.
In February 1969 the Joint Task Force published the Provisional Algal
Assay Procedure (PAAP). The PAAP was developed from the collective
knowledge and experience of persons who had fundamental knowledge of
algal physiology, algal growth responses, and experience with algal
assays of various types. It was determined that the PAAP should consist
of three fundamental test procedures: a Bottle Test, a Continuous-Flow
Chemostat Test, and an In situ Test. It was fully recognized that the
PAAP was tentative and that a great deal of research would be necessary
to sharpen each of the three procedures to determine their capabilities
and to compare test performance one with the other.
Shortly after publication of the PAAP, a group consisting of govern-
ment, university and industrial laboratories undertook a comprehensive
research program to improve and evaluate it. The program was coordinated
by the National Eutrophication Research Program of the Federal Water
Quality Administration located at the Pacific Northwest Water Laboratory,
-------
The first phase was concerned with comparing the Bottle Test and
"Continuous-Flow" Chemostat Test for assaying the algal growth-nutrient
concentration relationships in natural and enriched waters. Each
laboratory followed the same research plan for the evaluation using
algal test species from a common source. When specific problems were
recognized, certain laboratories were assigned to investigate them
after all laboratories had agreed to the research plan to be followed.
In inter!aboratory precision tests, using the Bottle Test, excellent
agreement in the data was obtained by the participating laboratories.
After nearly two years of research it is now felt that the Bottle Test
had undergone sufficient evaluation and refinement to be considered
reliable. As a result this document, the Algal Assay Procedure: Bottle
Test, was developed.
Publication of this document does not imply that the Continuous-Flow
Chemostat Test or the In situ Test, are no longer a part of the PAAP,
Both are considered important procedures, but further research must be
conducted in order to evaluate and refine them before they become ready
for universal use. Neither should it be implied that, because the
Bottle Test has been extensively tested, future changes or additions
will not be incorporated into it. Oust as with all procedures, it will
very likely be updated from time to time. Although the laboratory
procedure itself is ready for routine use, further evaluation relating
to several specific field situations will be necessary. Also, this
procedure will serve as a basis of comparison for further evaluation
-------
3
2. Pri nciple
This algal assay is based on Liebiq's law of the minimum which
states that "growth is limited by the substance that is present in
minimal quantity in respect to the needs of the organism." The test
in its present form is intended primarily for use in the following
general situations.
1. Assessment of a receiving water to determine its nutrient
status and sensitivity to change.
2. Evaluation of materials and products to determine their
potential effects on algal growth in receiving waters.
3. Assessment of effects of changes in waste treatment
processes on receiving waters.
Bottle algal assays consist of three steps, (1) selection and
measurement of appropriate parameters during the assay (for
example, biomass indicators such as total cell carbon), (2) presen-
tation and statistical evaluation of the measurements made during
the assay, and (3) interpretation of the results with respect to
the specific problem being investigated. Specifically, it is
intended that the test be used: (1) to identify algal growth-
limiting nutrients; (2) to determine biologically the availability
of algal growth-limiting nutrients; and (3) to quantify the bio-
logical response to changes in concentrations of algal growth-
limiting nutrients. These measurements are made by adding a
selected test alga to the test water and determining algal growth
-------
4
The test may also be used to determine whether or not various
compounds or water samples are toxic or inhibitory to algae. In
this connection caution should be observed when interpreting results
when there is little or no growth response in samples where sufficient
nutrients appear to be or are, in fact, present. The presence of
toxicants can prevent or inhibit algal growth even when nutrients
are not limiting.
It should be pointed out that test flasks are normally incubated
to facilitate free gas exchange at the air-water interface. Therefore,
since atmospheric carbon dioxide is available, the test as outlined
cannot be used to demonstrate algal growth limitations due to lack
of carbon in the water. The test, however, can be modified to obtain
such information, but more research is necessary.
2.1 Maximum specific growth rate is related to the concentration
of the rate limiting nutrient present. Maximum standing crop is
proportional to the initial amount of the limiting nutrient available.
2.2 Growth response - All comparative growth, responses should
be analyzed statistically and significant levels of the differences
reported. For most purposes a 95 percent significance level can be
considered statistically significant.
3. Planning and Evaluation of Algal Assays
The specific experimental design of each algal assay must be
dictated by the actual situation. It is extremely important that
all pertinent ecological factors be considered in the planning of
-------
5
obtained. As a minimum, the following specifics must be considered
by each investigator who plans to conduct algal assays for the
purposes listed above.
3.1 Selection of test waters - Water quality may vary greatly
with time and with location in lakes, impoundments and streams.
Sampling programs should be established so that meaningful data
wi11 be obtai ned.
3.11 Spatial variations - In a stratified lake or
impoundment it may be of value to sample both epilimnetic and
hypolimnetic waters. The use of transection lines are helpful in
sampling; samples from a transection can be taken from major depth
zones. In rivers and streams useful information may be obtained
by taking samples upstream and downstream from suspected nutrient
sources and from tributary streams. When materials are evaluated,
samples from a number of natural waters having a range of representa-
tive water qualities must be included.
3.12 Temporal variations - The nutrient content of natural
and waste waters often vary greatly with time. The variation may
not only be seasonal, but hourly. When sampling, this must be taken
into consideration.
3.2 Determination of limiting nutrients - Any of the essential
nutrients may limit algal growth. Bioassays can be designed to examine
in detail only a few nutrients which by preliminary testing have
been shown to be most likely limiting or in short supply. An example
-------
3.3 Evaluation of materials - In planning a program for evalua-
ting a material for its potential effect on receiving waters the
following factors must be considered:
(1) The distribution of the test material; e.g., local,
regional, national.
(2) The level of test material to be used; e.g., usage
in product or other measure of material involved.
(3) The chemical and/or physical nature of the material;
e.g., its theoretical potential for direct nutrient enrichment or
of indirect effects.
(4) The fate of the material; e.g., its chemical change
or biological degradation in waste treatment or surface waters.
(5) The pathway by which the material will reach the
receiving water; e.g., consideration of including the test material
in an appropriate range of waste treatment effluents or as a
component of surface runoff.
(6) The dilution factor for the stream receiving wastes
containing the test material; e.g., an appropriate range of dilutions
for introduction to receiving waters.
(7) The selection of appropriate test water; e.g., use
of appropriate test water for a material of local interest, but use
of full range of water qualities for material with broad distribution.
-------
7
3.4 Assessment of waste treatment processes - The algal assay
may be used to determine the algal growth stimulatory effects of a
given process effluent. When the assay is used for this purpose,
the overall evaluation must include consideration of the following
aspects:
(1) Typical and atypical conditions, nationwide and/or
local, under which the type of process effluent may enter the
environment.
(2) Growth parameter(s) and test organism(s) to be used.
(3) Which nutrient(s) is (are) limiting growth?
(4) Is there a change in the growth limiting nutrient as
a result of the process?
(5) What is the overall effect of a process or process
change?
An example of one possible treatment evaluation is
shown in Appendix 3.4.
4. Apparatus
4.1 Sampling and sample preparation.
4.11 Water sampler - non-metallic.
4.12 Sample bottles - autoclavable (such as borosilicate
glass, linear polyethylene, polycarbonate, or polypropylene).
4.13 Membrane filter apparatus - for use with 47 mm pre-
filter pads and an 0.45y porosity filter.
4.14 Autoclave or pressure cooker - capable of producing
15 psi (1.1 kg/cm2) at 250° F (121° C).
-------
8
4.21 Culture vessels - Erlenmeyer flasks of good quality
borosilicate glass such as Pyrex or Kimax. The same brand of glass
should be used within laboratory. When trace nutrients are being
studied special glassware, such as Vycor, polycarbonate, or coated
glassware should be used.
While the flask size is not critical, due to carbon dioxide
limitation, the surface to volume ratios are. The recommended
surface to volume ratios are as follows:
40 ml of sample in 125 ml flask
60 ml of sample in 250 ml flask
100 ml of sample in 500 ml flask
It is desirable to number permanently test flasks in order
that anomalous growth which appears to be related to specific flasks
can be identified arid those flasks eliminated from future tests.
4.22 Culture closures - foam plugs, loose fitting
aluminum foil or inverted beakers must be used to permit good
gas exchange (see Sec 9.2) and prevent contamination. Each
laboratory must determine for each batch of closures purchased
whether that batch has any significant effect on the maximum
specific growth rate and/or the maximum standing crop.
4.23 Constant temperature room, or equivalent incu-
bator, capable of providing temperature control at 24+2° C.
4.24 Illumination - "Cool-White" fluorescent lighting
to provide 400 ft-c (4304 lux) +10 percent or 200 ft-c (2152 lux)
-------
9
4.25 Light meter - several types may be used, but must
be calibrated against a standard light source or light meter (see
Appendi x 8).
4.3 General
4.31 Balance, analytical, capable of weighing 100 gm with
a precision of +0.1 mg.
4.32 Microscope - good quality general purpose.
4.33 Microscope illuminator - good quality general purpose.
4.34 Hemacytometer or plankton counting slide.
4.35 pH meter - scale of 0-14 pH units with accuracy of
+0.1 pH unit.
4.36 Oven, dry heat capable of temperatures of 120° C.
4.37 Centrifuge - capable of relative centrifugal force
of at least 1,000 x £.
4.38 Spectrophotometer or colorimeter - for use at 600-750my
4.4 Optional
4.41 Electronic cell counter.
4.42 Fluorometer.
4.43 Shaker table, capable of 100 oscillations per minute.
5. Sample Collection, Transport, Preparation and Storage
5.1 Collection - Use non-metallic water sampler and autoclavable
storage containers (see Sections 4.11 and 4.12). Containers should
not be re-used when toxic or nutrient contamination is suspected.
5.2 Transport conditions - Leave a minimum of air space in
transport container, keep in dark and at ice temperature.
5.3 Preparation - To enable the use of unialgal test species
-------
This "removal" necessitates either the separation or destruction of
the indigenous algae in the sample. Filtration and autoclaving are
recommended and the use of one or both depends upon the type of infor-
mation being sought. The effects of some pretreatment methods are
shown in Appendix 5.3.
5.31 Membrane filtration - should be used when it is desired
to remove indigenous algae to determine growth-limiting soluble nutrients,
which have not been taken up by filterable organisms, or in order to
predict the effect of adding nutrients to a test water at a specific
time. Pretreat 0.45y porosity membrane filter by passing at least
50 ml of distilled water through it. Discard filtrate. Then filter
quantity of the sample as needed under reduced pressure of 0.5 atmos-
phere or less. If there is a large amount of suspended material in
the sample, filtration through the 0.45y porosity filter pad should
be preceded by filtration through an appropriate filter (for example,
glass fiber) which is also pretreated as described above.
5.32 Autoclaving - should be used when it is desired to
determine the amount of algal biomass that can be grown from all
additional nutrients in the water, including those contained in
filterable organisms, which can be solubilized by autoclaving. The
sample should be autoclaved at 15 psi (1.1 kg-cm^) at 250° F (121° C).
The length of time of autoclaving will depend on the volume of the
sample, e.g., 30 min or 10 min/1, whichever is longer. After auto-
claving and cooling the sample should be allowed to equilibrate either
in an air or carbon dioxide atmosphere in order to restore the carbon
-------
11
level (it will generally rise on autoclaving). If an electronic particle
counter is to be used for all counting, the carbon-dioxide equilibrated
sample should then be passed through an 0.45y membrane filter.
5.4 Storage - Although it is known that changes do occur in
water samples during storage regardless of storage conditions, the
extent or chemistry of these changes is not well defined, and there-
fore attempts should be made to minimize the duration of storage.
Changes in samples should be minimized by keeping samples cool, in the
dark, using proper containers, and avoiding air spaces over sample.
Temporary storage prior to sample preparation should be in the
dark at 0-4° C. If prolonged storage is anticipated, the sample
should be prepared first and then stored in the dark at 0-4° C.
6. Synthetic Algal Nutrient Medium
6.1 Final concentration of nutrients.
6.11 Macronutrients - The following salts, Biological or
Reagent grade, in mi 11igrams per liter of glass-distilled water.
Compound
Concentration (mq/1)
Element
Concentration
NaN03
25.500
N
4.200
k2hpo4
1.044
P
0.186
MgCl 2
5.700
Mg
2.904
MgS04 7H20
14.700
S
1.911
CaCl2 2H20
4.410
C
2.143
NaHC03
15.000
Ca
1.202
Na
11.001
K
-------
12
6.12 Micronutrients - The following salts, Biological or
Reagent grade, in micrograms per liter of glass-distilled water.
Compound
Concentration (yg/1)
Element
Concentr
H3B03
185.520
B
32.460
MnCl2
264.264
Mn
115.374
ZnC 12
32.709
Zn
15.691
CoCl2
0.780
Co
0.354
CuCl 2
0.009
Cu
0.004
Na2MoO^
2H20
7.260
Mo
2.878
FeCl 3
96.000
Fe
33.051
Na2EDTA
2H20
300.000
6.2
Stock
solutions.
6.21 Macronutrients - Stock solutions of individual salts may
be made up in 1000 times the final concentration.
6.22 Micronutrients - The trace metals, FeCl3 and EDTA are
combined in a single stock mix at 1000 times final concentration.
6.3 Preparation of medium
6.31 Combination of stock solutions - 1 ml of each of the
stock solutions (6.21 and 6.22) is added to glass-distilled water to
give a final volume of 1 liter. The trace metal - FeC13EDTA mixture
(6.12) is added after filtration.
6.32 Pretreatment of uninoculated reference medium - For some
work, sterilization may not be required for experiments to be carried
out with freshly prepared culture media since the recommended assay species
-------
13
commensal (non-parasitic) bacteria. Stock cultures, however, should be
maintained in previously sterilized culture medium. It is recommended
that uninoculated sterile reference medium be stored in the dark to
avoid any (unknown) photochemical changes.
6.33 Prolonged storage - Reference medium for stock cultures
may be filter-sterilized as in 5.31 or autoclaved as suggested in 5.32.
7. Inoculum
7.1 Test algae (see Appendix 7.1 also).
7.11 Selenastrum capri cornutum Printz.
7.12 Microcystis aeruginosa Kutz. emend Elenkin (Anac.ystis
c.yanea Drouet and Daily).
7.13 Anabaena flos-aquae (Lyngb.) De Brebisson.
7.14 Diatom - not yet selected. If one is used add 10 mg Si/1
(101.214 mg NagSiOg 9H20/1) to the culture
medi um.
7.2 Source of test algae - Available from National Eutrophication
Research Program, Pacific Northwest Water Laboratory, Environmental
Protection Agency, 200 S.W. 35th Street, Corvallis, Oregon 97330.
7.3 Maintenance of stock cultures.
7.31 Medium - see 6.
7.32 Incubation conditions - 24 + 2° C under continuous
cool-white fluorescent lighting - 400 ft-c (4304 lux) + 10 percent for
S_. capri cornutum and 200 ft-c (2152 lux) + 10 percent for M. aeruginosa
-------
14
7.33 First stock transfer - Upon receipt of the inoculum
species, a portion should be transferred to the algal culture medium
in 6. (Example: 1 ml of inoculum in 100 ml of medium in 500 ml
Erlenmeyer flask).
7.34 Subsequent stock transfers - A new stock transfer
using an aseptic technique should be made as the first operation
upon opening a stock culture. The volume of the transfers is not
critical so long as enough cells are included to overcome significant
growth lag. A routine stock transfer schedule, such as weekly, is
recommended as a means of providing a continuing supply of "healthy"
cells for experimental work.
7.35 Age of inoculum - Cultures, one to three weeks old,
may be used as a source of inoculum. For Selenastrum a one-week
incubation is often sufficient to provide enough cells. The blue-
green species require a longer time to achieve maximum crop than does
Selenastrum. This slower growth of the blue-green assay species
should be considered in planning for sufficient inoculum to carry out
required experimental work. Thus, two to three weeks may be required
to provide inocula for assays with the blue-green species.
7.4 Preparation of inoculum - Cells from the stock culture should
be centrifuged and the supernatant discarded. The sedimented cells
should be resuspended in an appropriate volume of glass distilled water
containing 15 mg NaHCO^/l and again centrifuged. The sedimented algae
should again be resuspended in the water-bicarbonate solution and used
-------
15
7.5 Amount of inoculum - The cells suspended in the bicarbonate
solution (7.4) should be counted and pipetted into the test water to
give a starting cell concentration in the test waters as follows:
3
S^. capricornutum 10 cells/ml
3
M. aeruginosa 50 x 10 cells/ml
A. flos-aquae 50 x 10^ cell/ml
The volume of the transfer is calculated to result in the
above concentrations in the test flasks (Example: for S_. capri cornutum
5
5 x 10 cells/ml in the stock culture requires a 0.2 ml transfer per
100 ml of test water).
8. Test Conditions
8.1 Temperature. 24+2° C.
8.2 Illumination - continuous "cool-white" fluorescent lighting -
400 ft-c (4304 lux) + 10 percent for S_. capri cornutum and 200 ft-c
(2152 lux) + 10 percent for M. aeruginosa and A. flos-aquae. Intensity
is measured adjacent to the flask at the liquid level (see Appendix
8.2).
9. Procedure
9.1 Preparation of glassware - The recommended procedure is as
follows: All cylinders, flasks, bottles, centrifuge tubes and vials
are washed with detergent or sodium carbonate and rinsed thoroughly with
tap water. This is followed by a rinse with 10 percent solution by
volume of reagent hydrochloric acid (HC1); vials and centrifuge tubes
are filled with the 10 percent HC1 solution and allowed to remain a
-------
capacity with HC1 solution and swirled so that the entire inner surface
is bathed. After the HC1 rinse, the glassware is rinsed five times
with tap water followed by five rinses with deionized water.
Pipettes are placed in 10 percent HC1 solution for 12 hours or
longer and then rinsed at least 10 times with tap water in an automatic
pipette washer followed by a rinse with deionized water. Disposable
pipettes may be used to eliminate the need for pipette washing and to
minimize the possibility of contamination.
Cleaned glassware is dried at 105° C in an oven and is then
stored either in closed cabinets or on open shelves with the tops
covered with aluminum foil.
Before use, culture flasks are stoppered with plastic plugs or
covered with aluminum foil and autoclaved at 15 psi for 15 minutes.
Following autoclaving the flasks are prerinsed with the type medium
to be used for subsequent culturing and placed for 20-30 minutes, inverted,
on absorbant paper to drain.
9.2 pH Control - In order to insure the availability of carbon
dioxide the pH should be maintained below 8.5. This can be accomplished
by (1) using optimum surface to volume ratios; (2) by continuously
shaking the flask (approximately 100 oscillations per minute); (3)
by ventilation with air or air/carbon dioxide mixture; and, in extreme
cases, by (4) bubbling air/carbon dioxide mixture through the culture.
9.3 Growth parameters - Two parameters are used to describe the
growth of a test alga in the Bottle Test: maximum specific growth rate
and maximum standing crop. Either or both of these parameters may be
-------
17
9.4 Maximum specific growth rate
9.41 The maximum specific growth rate (ym,u) for an individual
ilia X
flask is the largest specific growth rate (p) occurring at any time
during incubation. The umax for a set of replicate flasks is determined
by averaging nmax of the individual flasks.
The specific growth rate, y, is defined by
ln(X2/X,) ,
u = -i 1 days
2 " tl
where = biomass concentration at end of selected
time interval
X, = biomass concentration at beginning of
selected time interval
tp-t, = elapsed time (in days) between selected
determinations of biomass
NOTE: If biomass (dry weight) is determined indirectly,
e.g., by cell counts, the specific growth rate may be
computed directly from these determinations without
conversion to biomass, provided the factor relating the
indirect determination to biomass remains constant for the
time period considered.
9.42 Laboratory measurements - The maximum specific growth
rate occurs during the logarithmic phase of growthusually between
day 0 and day 5and therefore it is necessary that measurements of
biomass be made at least daily during the first 5 days of incubation
to determine this maximum rate. Indirect measurements of biomass,
such as cell counts, will normally be required because of the difficulty
-------
18
The time at which measurements are made should be recorded for use in
the computations.
9.43 Computation of maximum specific growth rate - The
maximum specific growth rate (umax) can be determined by calculation
using the equation in Section 9.41 to determine the daily specific
growth rate (y) for each replicate flask and averaging the largest
value for each flask (see Appendix 11.31). It may also be determined
by preparing a semi-log plot of biomass concentration versus time
for each replicate flask. Ideally, the exponential growth phase can
be identified by 3 or 4 points which lie on a straight line on this
plot. However, the data often deviate somewhat from a straight line,
so a line judged to approximate most closely the exponential growth
phase is drawn on the plot. If it appears that the data describe two
straight lines, the line of steepest slope should be used. A linear
regression analysis of the data may also be used to determine the best
fit straight line. Two data points which most closely fit the line
are selected and the specific growth rate (u) is determined according
to the equation given in Section 9.41 (see Appendix 9.43 and Appendix
11.31). The largest specific growth rates for the replicate flasks are
averaged to obtain ymax«
9.5 Maximum standing crop
9.51 Definition - The maximum standing crop in any flask
is defined as the maximum algal biomass achieved during incubation.
For practical purposes, it may be assumed that the maximum standing
crop has been achieved when the increase in biomass is less than
-------
19
9.52 Laboratory measurement - After the maximum standing
crop has been achieved, the dry weight of algal biomass may be
determined gravimetrically using either the aluminum-dish or filtration
technique. For details of these techniques, see Appendix 9.61. If
biomass is determined indirectly, the results should be converted to
an equivalent dry weight using appropriate conversion factors.
9.6 Biomass monitoring - several methods may be used, but they
must always be related to dry weight.
9.61 Dry weight - gravimetrically (see Appendix 9.61).
9.62 By direct microscopic counting (hemacytometer or
plankton counting cell) or the use of an electronic particle counter.
A. flos-aquae, which is filamentous, is not amenable to counting with
an electronic particle counter. Microscopic counting can be facilitated
by breaking up the algal filaments with a high speed blender or by
sonication (see Appendix 9.62).
9.63 Absorbance - with a spectrophotometer or colorimeter
at a wavelength of 600-750 my. In reporting the results, the instrument
make or model, the geometry and path length of the cuvette, the wave
length used, and the equivalence to biomass should be reported (see
Appendix 9.63).
9.64 Chlorophyll - after extraction or by direct f1uorometric
determination (see Appendix 9.64). The equivalence between chlorophyll
content and biomass should be reported.
9.65 Total cell carbon - by carbon analyzer. Equivalence
-------
20
10. Spi kes
The quantity of cells that may be produced in a given medium is
limited by the substance that is present in the lowest relative
quantity with respect to the needs of the organism. If a quantity of
the limiting substance were added to the medium, cell production would
increase until this additional supply was depleted or until some other
substance became limiting to the organism. Additions of substances
other than that which is limiting would yield no increase in cell
production. Nutrient additions may be made singly or in combination,
including waste waters, and the growth response may be compared
to unspiked controls to identify those substances which limit growth
rate or cell production. The selection of spikes, e.g., nitrogen,
phosphorus, iron, sewage effluents, etc., will depend on the answer
being sought.
In all instances the volume of a spike should be as small as
possible. The concentration of spikes will vary and must be matched
to the waters being tested. Two considerations should be taken into
account when selecting the concentration of spikes: (1) the
concentration should be kept small to minimize alterations of the
sample, but at the same time it should be sufficiently large to yield
a potentially measurable response; and (2) the concentration of spikes
should be related to the fertility of the sample. To assess the
effect of nutrient additions, they must be compared to an unspiked
control of the test water. If the control is quite fertile, cell
production will be high and flask-to-flask variations in the controls
-------
21
10.1 Auxiliary spikes - In addition to spikes for the purpose
of determining stimulatory or inhibitory effects on algal growth in test
waters, it is sometimes necessary to check for the possibility that the
test water contains some toxic material which could influence results.
To check for toxic materials, the test water may be spiked with
the elements in complete synthetic medium (see Section 6). If no increase
in growth occurs, the presence of toxic materials is suspected.
In some situations, dilution of the sample or the addition of a
chelator will eliminate toxic effects.
11. Data Analysis and Interpretation
11.1 Introduction - The fundamental measure used in the Algal
Assay: Bottle Test to describe algal growth is the amount of suspended
solids (dry weight) produced and determined gravimetrically. Other
biomass indicators such as those shown in 9.6 may be used; however,
all results presented must include experimentally determined conversion
factors between the indicator used and the dry weight of suspended
solids. Results should be expressed in the units shown in Appendix
9.6. Several different biomass indicators should be used whenever
possible because biomass indicators may respond differently to any
given nutrient limiting condition.
11.2 Reference Curves - Results of spiking assays should be
presented together with the results from two types of reference
samples; the assay reference medium and unspiked samples of the water
under consideration. Preferably the entire growth curves should be
presented for each of the two types of reference samples. The results
-------
22
specific growth rate (with time of occurrence) and maximum standing
crop (with time at which it was reached).
11.3 Maximum specific growth rates - See Section 9.43.
11.31 Identification of growth rate limiting nutrients by
single nutrient spikes - Growth rate limiting nutrients can be deter-
mined by spiking a number of replicate flasks with single nutrients,
determining the maximum specific growth rate for each flask, and comparing
the averages by a Student's t-Test, or other appropriate statistical
tests (see Appendix 11.31).
11.32 Identification of growth rate limiting nutrients by
spiking with many nutrients - Data analysis for multiple nutrient spiking
can be performed by analysis of variance calculations. It is important
in multiple nutrient spiking to account for the possible interaction
between different nutrients; such interactions can readily be accounted
for by means of the above mentioned factorial analysis. An example
of an assay to determine the growth limiting nutrient is described
in Appendix 11 together with a factorial analysis computer program.
11.4 Maximum standing crop - See Section 9.5.
11.41 Identification of growth limiting nutrient - The
same methods which were described in Sections 11.31 and 11.32 for finding
the growth rate limiting nutrient can be used to determine the nutrient
limiting growth of the maximum standing crop (see Appendix 11.41).
11.42 Available concentration of growth limiting nutrient -
The "Available Concentration" of the growth limiting nutrient can be
determined by comparing the maximum standing crop in an unspiked sample
with the maximum standing crop in the reference medium having varying
-------
11.5 Confidence intervals - Both the maximum specific growth
rate and the maximum standing crop should be presented with the
confidence interval indicated. The calculation of confidence interval
for the average values presented should be based on at least five samples.
Consequently, a minimum of five replications should be made the
first time when an unfamiliar source water is analyzed. The results
of these five replicates are then used to calculate the standard
deviation. Subsequent samples from the same source can be analyzed
using only three replicates and reported with the confidence interval
established for that source water (see Appendix 11.5).
11.6 Rejection of outliers - When algal assays are conducted
it is often observed that one of the flasks among replicates shows a
growth difference from the remainder of the replicates. Such outliers
can be eliminated from the results if they fall outside the limits
indicated in Appendix 11.6. A laboratory should keep track of individual
flasks which result in outliers. Flasks which produce outliers more
than once should not be used in further algal assays.
11.7 Evaluation of assay results - The overall evaluation of
assay results consists of two parts. The first part is the determination
whether a given assay result is significant when considered as a
laboratory measurement. Several methods are available such as the
Student's t and Analysis of Variance techniques presented in Appendix
11. However, it must be emphasized that there is as yet no unique
method available to determine significant responses. Each experimental
evaluation must be designed based upon the specific objectives using
-------
The second part of the overall evaluation is the correlation of
laboratory assay results to effects observed or predicted in the field.
This phase is underway but no specific guidelines are yet available;
investigators should note the general considerations presented in Section
-------
APPENDICES*
*Each Appendix is numbered to correspond with that Section to which its
-------
APPENDIX 3
3.2 Determination of limiting nutrients - This appendix is
included to provide illustrations of how the technique of spiking
may be used to conduct bioassays. The examples are included not
necessarily to indicate how bioassays should be conducted, but rather
to illustrate some of the rationale and logic that may be appropriate
in certain instances. Depending upon the objectives, bioassays may
be more elaborate or simpler than the examples given.
Example 1 - Spiking with One Nutrient
Assume that a sample has been collected from a relatively
infertile lake, treated to remove indigenous organisms, and a bioassay
is to be conducted to obtain answers to the following:
1. Which nutrient limits maximum cell production in
the sample?
2. What is the biologically available concentration
of that nutrient?
3. Is the sample sensitive to changes in the concentration
of the limiting nutrient?
Assume further that preliminary testing has shown that nitrogen and
phosphorus are the only two potentially limiting nutrients, and toxic
materials are not present in the sample.
To achieve the stated objectives the following experimental
-------
28
TABLE 1
Experimental Design, Example 1
Treatment No. Flasks
1. Lake Water Control 3
2. Phosphorus Spikes
Lake Water + .005 mg P/l 3
Lake Water + .015 mg P/l 3
Lake Water + .050 mg P/l 3
3. Nitrogen Spikes
Lake Water + .075 mg N/1 3
Lake Water + .225 mg N/1 3
Lake Water + .750 mg N/1 3
4. Combined Spikes
Lake Water + .005 mg P/l + .075 mg N/1 3
Lake Water + .015 mg P/l + .225 mg N/1 3
Lake Water + .050 mg P/l + .750 mg N/1 3
5. Growth References - Phosphorus
(Medium*-P) 3
{Medium -P) + .005 mg P/l 3
(Medium -P) + .015 mg P/l 3
(Medium -P) + .050 mg P/l 3
6. Growth References - Nitrogen
(Medium*-N) 3
(Medium -N) + .075 mg N/1 3
(Medium -N) + .225 mg N/1 3
(Medium -N) + .750 mg N/1 3
7. Ful 1-Strength Medium _3^
Total 57
~Medium refers to the synthetic algal nutrient medium (see Section 6.).
(Medium -P) refers to the medium prepared without phosphorus.
-------
29
1. Lake Water Control - Lake water blanks provide the basis for
comparison of the other treatments and provide a measure of the general
fertility of the sample. (In some instances, increased replication of the
control would be desirable.)
2. Phosphorus Spikes - This series is included to determine if the
sample is sensitive to additions of phosphorus. Since the sample was
thought to be infertile, the concentrations of the spikes are relatively
low. Three separate concentrations were selected to increase the possibility
of obtaining a measurable response if, in fact, phosphorus was the limiting
nutrient, and also to characterize the response of the sample to additions
of phosphorus.
3. Nitrogen Spikes - Rationale same as for phosphorus spikes.
4. Combined Spikes - This series is included to investigate the
possibility that both nitrogen and phosphorus were present in amounts so
small that additions of these nutrients singly would not yield a measurable
response.
5. Growth References - Phosphorus - The controls are included to:
a) obtain a measure of "nutrient carry-over" in the inoculum (medium -P
flasks); b) establish that the inoculum is quantitatively responsive to
changes in the concentration of phosphorus; and c) provide a relative scale
for quantifying the amount of phosphorus available to the test organism.
6. Growth References - Nitrogen - Rationale same as for phosphorus
controls.
7. Full-Strength Medium - This treatment is included to provide a
general check on cell growth and to provide an index for comparison to other
-------
30
3.3 Evaluation of Materials
Example (X)
Hypothetical Situation: Assume that a new household material
will be introduced into a regional market, that the material will
enter surface waters through the sewage system equipped with secondary
treatment, that expected levels in the sewage will be 1 mg/1, and
that there are two large lakes in the regional drainage basin. The
lakes have in the past supported populations of green and blue-green
algae and, therefore, Selenastrum capricornutum and Microcystis
aeruginosa were selected as test organisms.
A test protocol for evaluating the effect of the new material
on algae in the lake waters might be:
1. Blank Lake Water
New Material Spikes
2. Lake water + New Material at 0.1 mg/1
3. Lake water + New Material at 1.0 mg/1
4. Lake water + New Material at 10.0 mg/1
Sewage Spikes
*5. Lake water + primary effluent without New Material
*6. Lake water + primary effluent containing 10 mg/1 New Material
7. Lake water + secondary effluent without New Material
8. Lake water + secondary effluent fed with primary effluent
containing 10 mg/1 of New Material
*Assuming a significant portion of primary effluent bypasses secondary
treatment. Each variable is run in both lake waters. Sewages are
added at a concentration of 1-2 percent by volume. Three replicates
are run on each variable (3 reps, x 11 variables x 2 waters x 2
-------
31
Nutrient Equivalent Spikes
9. Lake water + major nutrient (e.g., P) equivalent to that
contained in 10 mg/1 New Material
10. Lake water + major nutrient (e.g., N, etc.) equivalent to that
contained in 10 mg/1 New Material
11. Lake water + 10 percent medium* (as indicator of a possible
toxicant present in lake water)
12. Medium - Optional (to make certain inoculum is viable)
3.4 Example of method to assess waste treatment processes -
This example illustrates one possible application of the algal assay
in the evaluation of the potential treatment process. The process
evaluation described below was made on a process which consisted of
percolation of a secondary effluent through approximately 400 ft of
natural gravel.
The experimental design was based on local conditions where
the process effluent would be discharged. In this case, the process
effluent would be discharged directly into artificial lakes without any
appreciable natural dilution. Consequently these tests were conducted
on undiluted process effluent.
The maximum standing crop was chosen as the parameter to be
evaluated based on the conditions of very long hydraulic residence
times in the lakes which would receive the effluent. Selenastrum
capricornutum Printz was selected as the test organism because
unicellular green algae had been found to predominate in the lakes
in previous years.
*Medium refers to synthetic algal culture medium (see Section 6).
Library
Pacific Northwest Water Laboratory
EPA, WQO ^
200 S. W. 35th Street
-------
32
The limiting nutrients were determined in two steps for both
the influent to and effluent from the percolation grounds.
The first step consisted of dividing the possible growth
limiting nutrients into four groups:
I Trace metals
II Iron and Manganese
III Phosphorous and Nitrogen
IV Macronutrients (Ca, Mg, CI, etc)
These four groups were then added alone and in all possible
combinations to samples of the influent and effluent and the maximum
standing crop determined. Two replicates should be made for each of the
combinations tested (i.e., a total of 32 tests for each water tested).
A typical factorial experiment layout is shown in Table 1.
The results, cell counts, and dry weights can be analyzed as a 2x2x2x2
factorial experiment. A computer program for analyzing such a factorial
-------
TABLE 1
DESIGN FOR 24 FACTORIAL EXPERIMENTS
TO DETERMINE LIMITING NUTRIENTS
FACTORS
30% *
Macronutrients
added
+
.
30% *
P & N added
+
-
+
-
30% *
Fe & Mn added
+
-
+
-
+
-
+
-
30% *
other trace
metals added
+
+
_
+
+
_
+
.
+
+
.
+
Rep I Flask No.
1
3
5
7
9
11
13
15
17
19
21
23
25
27
29
31
Rep II Flask No.
2
4
6
8
10
12
14
16
18
20
22
24
26
28
30
32
"Factor" in
print out
15
13
14
10
12
9
7
4
11
8
6
3
5
2
1
-------
35
APPENDIX 5
5.3 Sample preparation - The ability of filtration and heat
treatment techniques to remove "indigenous" algae from culture media
as well as natural water samples enriched with algae was evaluated.
After treatment, samples were incubated under standard Provisional
Algal Assay Procedure batch conditions (without inoculation to
detect growth of resistant algae). The change in chemical composition
of the samples after treatment is shown in Table 1. Filtration through
Whatman No. 41 filter paper removed up to 67 percent of the total
phosphorus. Membrane filtration caused no apparent change in the
COD or total phosphorus of an oligotrophic water but removed close
to 67 percent of the total phosphorus of eutrophic water samples
and from 22 to 71 percent of the Fe of samples. Centrifugation also
removed substantial quantities (up to 55 percent) of the total
phosphorus of samples.
Pasteurization, autoclaving and sonication obviously do
not remove constituents from the water samples unless these are lost
under the conditions used, e.g., volatile substances driven off or
carbon dioxide evolution. However, total carbon and total phosphorus
analyses of autoclaved and raw samples of eutrophic water showed no
essential differences, although reductions of total phosphorus
concentration in other eutrophic waters by autoclaving and pasteuri-
zation have been observed. These decreases were not consistently
observed and were generally below 20 percent; however, in one
-------
36
Analyses of Ca, Na, K, Mg, CI, SO^, and NOg showed that
filtration and heat treatment did not significantly affect the concen-
trations of these ions as shown in Table 2. Small changes were observed
but these could have been due to the precision of analyses. However,
autoclaving appears to cause an increase in the soluble silica concen-
tration, possibly from the glass containers.
TABLE 1
CHANGE OF COD AND TOTAL PHOSPHORUS CONTENT OF SAMPLES
INDUCED BY FILTRATION AND CENTRIFUGATION TECHNIQUES
Change in Concentration
Range in Percent
Pretreatment
COD
Total
Phosphorus
Fi1tration
(Whatman No. 41)
H.D.
5.0 to 66.9
Fi1tration
(0.45y to 0.5y membranes)
0.0a to 72.3b(4) 0.0a to 66.7b
Centrifugation
N.D.
13.8 to 54.9
a01igotrophic water
bEutrophic water
The effect of pretreatment on maximum standing crop and
-------
37
TABLE 2
EFFECT OF PRETREATMENT METHODS ON CHEMICAL
CONSTITUENTS IN SAMPLES
Characteristic
Raw Water
Nutrient
Concentration
c0-mg/i
Pretreatment and Terminal Concentrations
F
A
FA
S
FS
Ct-mg/1
Ct-mg/l
Ct-mg/l
Ct-mg/l
Ct-mg/l
SO*
<10.0
<10.0
<10.0
<10.0
<10.0
<10.0
Ca
8.0
5.1
11.5
9.6
10.8
8.0
5.1
11.3
10.4
10.7
9.8
8.4
8.2
Na
6.1
6.6
6.4
6.6
6.3
6.8
K
0.7
0.6
0.6
0.6
0.6
0.6
CI
4.0
5.0
4.0
4.0
4.0
5.0
nh3
0.0
0.0
0.1
0.6
0.1
0.0
0.1
1.2
Urn v _
NO 3
<.1
<.1
<.1
<.1
<.1
<.1
<.1
<.1
Mg
.5
3.8
5.4
4.1
.4
3.8
5.8
4.1
3.2
3.2
3.3
Soluble
Si
1.8
1.8
5.9
6.3
1.9
1.8
F = Filtered through 0.45 to 0.5u pore size membranes
A = Autoclaved S = Sonicated
-------
38
TABLE 3
THE EFFECT OF PRETREATMENT ON THE MAXIMUM STANDING CROP
Relati ve
Maximum Cell
Product!on
Response
Samples
Bantam
Lake
Wawayanda
Lake
Green
Pond
Pond
Water
Lake
Anza
"XF
2.5
3.4
/\
Xp
*F
--
--
2.5
*FW
XF
--
--
3.1
S\
^A_
XFW
1.1
A
XP
--
1.3
XFA
A
XF
0.37
1.7
0.39
_ _
>
XpF
A
XF
3.3
«. _
-
*AF
A
XF
15.7
17.7
5.6
0.5
--
XAF
XpF
5.3
¦"
~
XAF
XFA
42.3
3.3
1.3
mm »
-------
39
X maximum cell concentration
- not determined
A autoclaving
F filtration
FW Whatman filtration
P pasteurization
FA filtration followed
by autoclaving
PF pasteurization followed
by filtration
AF autoclaving followed
by filtration
TABLE 4
ALGAL GROWTH RATES FOR SEVERAL SAMPLE PRETREATMENTS
Treatment
Mean Maximum
Specific Growth Rate
Day-'
Coefficientuof
Variation
%
Autoclaved
0.46a
26.4
Pasteurized
0.33
43.0
Filtration Whatman
No. 41
0.25
24.0
Filtration Membrane
0.45 u
0.25
29.3
aMean growth rate of four replicate flasks.
-------
The maximum standing crop values of pasteurized samples were found
to be significantly (p >_ 0.95) higher than the values of membrane
filtered samples. Similar results were observed for Whatman No. 41
filtered and autoclaved samples when compared to membrane filtered
samples. However, the results were quite variable and not statistically
different (i.e., p < 0.90). Higher values were observed with heat
treated samples as compared to membrane filtered samples as shown in
Table 3.
Autoclaving of samples (whether followed by filtration or
not) resulted in higher maximum standing crop values than observed
with membrane filtered samples in all cases but one. The latter was
an oligotrophic water in which membrane filtration apparently did not
remove much particulate matter. Pasteurization also resulted in
values higher than those for membrane filtration; however, generally
the maximum standing crop of pasteurized samples was less than that
of autoclaved samples. Membrane filtration reduced the maximum
standing crop even more than that observed with Whatman No. 41 filtered
samples. The highest values were obtained when samples were autoclaved,
followed in order by pasteurization, Whatman filtration and membrane
fi 1 tration.
Autoclaved samples resulted in the highest maximum specific
growth rates with the pasteurized samples producing the next highest
maximum specific growth rate. The lowest maximum specific growth rates
were obtained in Whatman filtered and membrane filtered samples. The
maximum specific growth rate of the autoclaved samples was observed
to be significantly (p :> 0.95) higher than the maximum specific growth
-------
APPENDIX 7
7.1 Comments on the Taxonomy, Morphology, and Reproduction of
Algal Test Species.
It is recommended that three standard test organisms be used
in the algal bioassay procedures for bottle tests and continuous flow
chemostat tests, viz: Selenastrum capricornutum, Anabaena flos-aquae,
and Microcystis aeruginosa. These three species are recommended for
use because they provide a representative cross-section of the various
different types of algal species likely to be found in a variety of
waters of different nutritional status. Selenastrum is a unicellular
or loosely aggregated colonial organism of the green algae or Chlorophyceae.
The two remaining species are of the blue-green algae, or Cyanophyceae.
Anabaena flos-aquae is a filamentous organism in which heterocysts
occur; it is, therefore, a species capable of fixing nitrogen; Microcystis
is a unicellular or loosely-aggregated colonial organism in which
heterocysts do not occur and it is, therefore, not a nitrogen fixer.
The following comments on the structure, reproduction, and form of
the three entities are presented for the benefit of those who wish
to use the algal bioassay procedures but who are not fully acquainted
with the test organisms.
7.11 Selenastrum capricornutum
Selanastrum capri cornutum is a green alga, Chlorophyceae,
of the order Chlorococcales. Entitles attributed to this order are
characterized by their unicellular or colonial habit in which the cells
-------
42
or for the greater part of it. The systematic subdivision of the Chlorococcales
is dependent upon the mode of reproduction, the cellular shape and
characteristics of the colony. Selenastrum is characterized by the
shape of individual cells which are curved in the shape of a new moon.
These may occur as single cells or aggregated into groups, the cells
being attached to one another at their convex faces by a relatively
minute gelatinous mass. The cell or the aggregate of cells does not
possess a prominent gelatinous envelope. The cell groups may contain
2 to 12 cells and several groups frequently become attached to one
another to form cell masses containing 50 or more cells. The larger
cell masses are not normally found in actively agitated cultures although
they do occur frequently in field collections. There is usually a
single chloroplast which in young cells lies on the convex side whereas
in older cells it may appear to fill the cell completely. Some authors
have reported a single pyrenoid, which is usually associated with
starch formation in Chlorophyceae, whereas other authors have claimed
that no pyrenoid is present. There is considerable variation in the
size of individual cells, which may range from 10 to 48 y in length
and 3 to 9 y in breadth. The cell size appears to reflect the state
of nutrition and the rate at which cell division is occurring. Various
species of the genus have been described on the basis of cell size
although most of these would appear to be of very dubious validity.
The most frequent form of reproduction in Selenastrum capricornutum
is by a process of cell division. Cells give rise directly to two
or more cells of similar shape to the parent cell or colony; the daughter
-------
7.12 Microcystis aeruginosa
Microcystis aeruginosa is a blue-green alga, Cyanophyceae,
of the order Chroococcales. Entities attributed to this order are
characterized by their unicellular or colonial habit and occurrence
in mucilaginous masses which are commonly free-floating. The cells
are small, spherical or ellipsoidal, 1 to 6 y in diameter. As in
all Cyanophyceae, there are no chloroplasts as such, the photosynthetic
pigments being distributed generally throughout the cytoplasm of the
cell giving it a granular appearance. The cells rarely occur singly
but are usually aggregated into masses which are held together by a
common gelatinous matrix of an extremely watery consistency. In quiet
waters of ponds and lakes, these gelatinous masses may measure as
much as a centimeter in diameter and be of a spherical, elongate or
highly irregular shape. There are no heterocysts present and the
cells do not appear to be differentiated from one another in any way.
Reproduction appears to occur by division of the cells to give two
daughter cells and the fragmentation of the gelatinous masses with
age. There are no reports of any forms of special asexual or sexual
reproduction in Microcystis.
7.13 Anabaena flos-aquae
Anabaena flos-aquae is a blue-green alga, Cyanophyceae,
of the order Hormogonales. Entities attributed to this order are
characterized by their unbranched filamentous organization in which
heterocysts occur. The normal cells composing the filaments are spherical
or slightly elongate in shape, 4 to 10 y in length and 4 to 7 y in
-------
sheath. As in all Cyanophyceae there are no chloroplasts as such,
the photosynthetic pigments being distributed generally throughout
the cytoplasm of the cell to give it a granular blue-green coloration.
The filaments are usually straight or occasionally with a slight curvature
but relatively stiff and rigid. Occasionally, the filaments may aggregate
to form free-floating masses but this is extremely rare in culture.
In quiet waters of ponds and lakes such masses may measure as much
as a centimeter in diameter, floating on the surface as an obvious
bloom. Heterocysts are cells which are much larger than ordinary
vegetative cells and possess a well-developed thick wall. In Anabaena,
the heterocysts normally occur only in an intercalary position. They
measure 8 to 12 p in length and 5 to 11 p in breadth. The heterocysts
are formed by the conversion of vegetative cells; this conversion
involves the loss of the normal blue-green pigmentation and the development
of a highly refractive cytoplasmic content of yellowish color. With
age the cell contents of the heterocysts may disappear completely
and they then appear as empty structures. The most crucial characteristic
of the heterocysts is that they appear to be the sites of nitrogen
fixation. Reproduction in Anabaena, as in most Cyanophyceae, is a
process about which little is known. Vegetative fragmentation is
probably the most frequent method of multiplication. In Anabaena,
structures which are referred to in the literature as "spores" or
"akinetes" occur with regularity. These are of an elongate cylindrical
shape and much larger than the vegetative cells, measuring as much
as 60 to 80 u in length. They arise by the conversion of normal vegetative
cells which increase markedly in size and develop a thickened wall.
-------
45
considerable time before germinating to form a new filament. There
are no indisputable reports of sexual reproduction in Anabaena, or
-------
APPENDIX 8
8.2 Illumination - This appendix describes the effect of light
intensities on the three test organisms, Selenastrum capricornutum,
Anabaena flos-aquae, and Microcystis aeruginosa. The data in Fig. 1
were obtained using the medium of the Provisional Algal Assay Procedure
(PAAP), February 1969. Growth responses in the present medium (Sec.
6.) have not been tested under the same range of light intensities,
but any differences due to a different medium are expected to be
minor.
Plots of cell numbers show that the logarithmic growth phase
occurs normally within the first week for all three species (Fig. 1).
The growth rate of the green alga, Selenastrum, is almost double that
of the blue-greens, Microcystis and Anabaena. Within the range of
light intensities studied, the growth rate for all three species in-
creased with increasing light intensity.
In culturing these species, therefore, producing enough cells to
serve as a test inoculum may require a longer period of time for
blue-greens than for greens, e.g., 2 weeks vs. 1 week, if a proposed
test is of sufficient size to require a large number of cells.
The maximum standing crop, determined at the end of the three-
week growth cycle as algal dry weight, also increased with increasing
light intensity (Fig. 2). The relationship of maximum standing crop/
light intensity was not so obvious, however, when the algal production
-------
SEIENASTRUM
GIOWTM I
RATI. -
E
M
lil
V
20
24
DAYS
| i i i | i i i | i <i [ fri|iir-
MICXOCYSTIS
O SO ft c
D wo h < »
A MO ft c M
o 300 ft c .70
10* * ¦ ¦ 1 1 ' 1 ' ' 1 ¦ I i i « 1 ' i ' ' ' 1 '
12
DAYS
16
20 24
RATf.
0 40 fftc
~ 100 ft C
A 400 ft C
o 300 ft c
-------
SELENASTRUM
MICROCYSTIS
ANABAENA
250
200
cn
E
^ 150
z
O
m
^ 100
ae.
a
50
100 200 400
s. B S S
50 100 200 300
FOOT-CANDLES
50 100 200 300
-------
50
In the aforementioned light intensity study, a conclusion was
drawn that the three algal test species could be grown at one convenient
light level of 200 ft-c. Nevertheless, as practically all the PAAP
development work had been done with one organism, Selenastrum, at
the light level of 400 ft-c., it was not seen fit to make a change.
There also was evidence of a slight growth rate limitation of
Selenastrum at 200 ft-c., and it was decided not to introduce that
variable into the test. No data were collected at illumination levels
above 400 ft-c. to determine if light at 400 ft-c. could be limiting
for Selenastrum.
NOTE: Light meters used in the assay laboratory to
determine light intensities should be standardized prior
to use. Commercial laboratories can perform such
standardizations by comparison with calibrated lamps
supplied by the National Bureau of Standards.
Standardization should include both intensity and
color temperature comparisons if the light meter to be
used is designed for a color temperature different from
the color correlated temperature (4200°K) of the "Cool-
White" fluorescent light source specified in the assay
-------
APPENDIX 9
9.43 Computation of maximum specific growth rates (see graph
on next page).
9.6 Biomass monitoring
MEASUREMENT UNITS
Suspended solids
mg l"1
Absorbance (optical density)
uni ts * cm"
Suspended carbon
mg l"1
Cell counts
number ml
Cell volume*
y3 ml"1
Chlorophyll fluorescence
units cm"
Extracted chlorophyll
yg l_1
9.61 Dry weight - This method will be particularly useful
for assessing the growth of Anabaena flos-aquae. The cells of this
alga grow in filaments and it is difficult to obtain accurate cell
counts. The method may also be used with S_. capricornutum, M.
aeruqinosa, and other species of algae. In any case, however, it
should only be used with either relatively dense cultures or large
volumes of thinner cultures. Otherwise, the error may be large.
Two methods may be employed.
Method I -- A suitable portion of algal suspension is
centrifuged, the sedimented cells washed three times in distilled water
containing 15 mg NaHCOg/l, transferred to tared crucibles or aluminum
cups, dried overnight in a hot air oven at 105°C and weighed. This
-------
52
: i
- o
o ©
o
Time
Cell
Count
(days)
(cells/ml)
0
1.0
x 10J
1
2.1
x 103
2
5.8
x 103
3
2.5
x 104
4
1.7
x 105
5
3.3
x 105
6
6.3
x 105
8
2.3
x 1Q6
11
2.4
x 106
14
1.6
x 106
17
2.4
x 106
By inspection of the
plot, the maximum
growth rate occurred
between days 2 and 4.
Vimav - In(1.7x105/5.8xl03)
ilia X 2 "
= ln(29.3)
2
= 3.38/2
= 1.69 days"1
W \ t t 8 To 1*2 14 16 18 '
Time, days
-------
53
method is more sensitive than Method II, but is open to error through
loss of cells during washing.
Method II This method involves filtering a measured
(R\
portion of algal suspension through a tared Milliporev ; filter. The
filter recommended is type AA with an 0.80 micron pore size.
The method is as follows:
1. Dry filters for several hours at 90°C in an oven.
The filters may be placed in folded sheets of paper upon which the weights
or codes may be written.
2. Cool filters in a desiccator containing desiccant.
3. Filter a suitable measured aliquot of the culture
under a vacuum of 0.5 atmosphere. Normally 10 ml is sufficient, but
in thin cultures more may be required.
4. Rinse the filter funnel with 50 ml distilled water
using a wash bottle and allow the rinsings to pass through the filter.
This serves to transfer all of the algae to the filter and to wash
the nutrient salts from the filter.
5. Dry the filter in its paper folder at 90°C, cool
in desiccator, and weigh.
6. To correct for loss of weight of filters during
washing, wash two blank filters with 50 ml of distilled water, pouring
it through slowly under reduced vacuum. Dry and weigh filters and
record weight loss. This correction is not large, but is essential
for meaningful results on thin cultures. For example, if 10 ml have
been filtered and yield a difference between tare and final of 1.10 mg
and the blank has lost 0.02 mg, then the culture contains:
-------
NOTE: a. The drying temperature has been selected
to avoid damage to filters.
b. In weighing, do not hang the filters over
the edge of the pan or electrostatic attraction of the
filter to the balance will result in an error.
9.62 By direct microscopic counting - Apparent irregularities
in the Anabaena growth curves are often due to the problem inherent in
counting filamentous algae, i.e., the difficulty in breaking up the
filaments properly, without cell damage, in order to obtain a representa-
tive sample for counting. Filament-breaking techniques which have been
used with varying degrees of success include the use of (1) a syringe,
(2) an ultrasonic bath, (3) a high speed blender, and (4) vigorous
swirling with glass beads. While none of these techniques is without
drawbacks, expelling the sample forcefully through a syringe against
the inside of the flask is the most satisfactory way to break up tight
clumps of filaments. Other methods of biomass measurement such as dry
weight, absorbance, or chlorophyll fluorescence are often considered
to be more suitable than cell counts for growth assessment of filamentous
algae.
9.63 Absorbance - Absorbance or optical density, as
defined by Beer's Law Expression D = log i = log y0- = acl is usually
1 Lx
derived for absorption of light by molecules of solute in homogenous
solution. It can be derived also for a suspension of uniform particles
but with some necessary added restrictions. For particles of bacterial
-------
by virtue of scattering caused by cell reflections and refractions.
The fraction of the latter which reach the light measuring receiver
depends upon the instrument design. A large receiver close to the
cuvette catches much of the scattered light (i.e., is insensitive to
scattering). A small receiver far from the cuvette in a long-focused
or diaphragmed optical path catches very little scattered light {i.e.,
is very sensitive to scattering. Reference: Mastre, H., J. Bacteriology,
30:335 (1935).
Most current instruments are likely to be more sensitive
to scattering than to absorption as evidenced by effect of wavelength.
A simple test is the following: For a green alga, light absorption by
pigments in vivo wi11 show relative optical densities for 600: 680:
750 my ratios, such as 70: 500: 1 (ratios probably correct in order of
magnitude). In practical measurements, without elaborate precautions to
avoid effects of scattering, the ratios will always be very much less.
In any photometric measure of optical density or absorbance
considerations of precision lead to simple rule of thumb that measure-
ments be limited to a range of 0.05
-------
assurance that the relation between D and X will be constant and
independent of culture conditions. As noted above, the optical density
measured is a complex function of volume, size and pigmentation of the
cells. Hence, the relation between D and X should be examined on
different batches of algae which best simulate actual conditions of
the test.
9.64 Chlorophyll - The use of fluorescence to determine
phytoplankton chlorophyll (extracted and in vivo) has been reported
to be more sensitive and less cumbersome than the familiar trichometric
method (Yentsch and Menzel, 1963; Lorenzen, 1966).
Chlorophyll, like many organic molecules, possesses
the ability to fluoresce. Simply stated, fluorescence is the ability
of a substance to absorb light energy at one wavelength and emit
this energy at a longer wavelength. Yentsch (1963) reported maximum
absorption (excitation) of chlorophyll in an acetone solution at
430 my and maximum emission between 650 and 675 my.
Lorenzen (1966) reported that the fluorescence of in
vivo chlorophyll is considerably less efficient than dissolved chlorophyll,
yielding only about 1/10 as much fluorescence per unit weight as the
same amount in solution.
Strickland and Parsons (1965) state that many plant
cells resist complete extraction in acetone and certain species retain
50-90 percent of their pigments. The use of a sonic disintegrator
may have limited value in increasing extraction efficiency. The use
of a tissue grinder, as recommended by Yentsch and Menzel (1963), is
-------
57
but even this approach fails to give complete extraction in a reasonable
time with certain species.
The correlation of (in vivo) fluorescence units and cells
per milliliter of Selenastrum capricornutum (Figure 1) shows excellent
agreement during the logarithmic phase of growth, indicating the use-
fulness of fluorescence to evaluate productivity. It should be pointed
out that the chlorophyll ^content of algal cells grown in culture
medium, natural waters, and sewage effluents is not always constant
and is subject to change dependent upon the physiological condition
of the cells. Nitrogen deficient cells, for example, usually have a
lower chlorophyll ^content than non-nitrogen limited cells even
though the cell mass (cells/ml and dry weight) may be in close agree-
ment. However, in vivo (direct) fluorescence measurements can aid in
the evaluation of (valid) increases in cell counts attributed to
increased algal growth.
The following methods of fluorometric determination of
chlorophyll a in vivo (directly) and after solvent extraction are
recommended,
I. Equipment and Supplies.
Millipore filtration equipment designed to hold 47 mm diameter
membrane or glass filters. Stoppered graduated centrifuge tubes of
15-ml capacity having both glass and polyethylene stoppers and "small
volume" cuvettes having a cell path of 1.0 cm, but holding 10 ml or
less of solution.
Fluorome.ter capable of measuring chlorophyll, e.g., modified
Turner or equivalent. Tissue grinder. Arthur H. Thomas #4288-B or
-------
10
CD 4
110
z
3
u
o
Ld 3
O in3
CO
LJ
CC
O
3
w 10
§
-j
UJ
cr
10 =-
S. Copricornutum
Sept. 1970
J LI
i I
I I I 1 TT
I I I III-
I 1 I
10
10
10
10 10
CELLS/ml
-------
II. Reagents.
A. 90 Percent acetone: Distill reagent grade acetone over
about 1 percent of its weight each of anhydrous sodium carbonate
and anhydrous sodium sulphite. Collect fraction boiling at a constant
temperature near 56.5°C (uncorrected). Pipette 100 ml of distilled
water into a one-liter volumetric flask and bring up to mark with
acetone to contain exactly 1000 ml. Store tightly stoppered in a
dark glass bottle between use. This reagent can be conveniently
dispensed from a polyethylene wash bottle which should be kept
nearly full.
B. Magnesium carbonate suspension. Add approximately 1 gram of
finely powdered magnesium carbonate (light wt. or "Levis" grade) of
analytical reagent quality to 100 ml of distilled water in a stoppered
Erlenmeyer flask. Shake vigorously to suspend powder immediately
before use.
III. In vivo (direct reading) procedure.
A. Sampling
1. Swirl flasks to insure homogenous suspension of algal cells
2. Pipette cell suspension aliquot (5 ml minimum) into small
beaker or vial.
B. Measure Fluorescence
1. Zero fluorometer with a distilled water blank before
each sample reading and change in sensitivity setting.
2. Pour well-mixed sample into cuvette and read fluorescence.
If reading .(scale deflection) is over 90 units, use lower sensitivity
setting, e.g., 30 x >10 x >3 x >1; conversely, if reading is less than
-------
60
3. If samples fail to stay in range, dilute accordingly.
4. Record fluorescenceunits, based on a common sensitivity
factor, e.g., a reading of 50 @ 1 x = 1500 @ 30 x.
IV. Extraction (dissolved chlorophyll) procedure.
A. Filter measured sample under vacuum through a glass fiber
filter. Add - 1.0 ml magnesium carbonate suspension and drain filter
thoroughly under suction.
B. Place filter into the bottom of tissue grinding tube. Add
2 ml of 90 percent acetone to the grinding tube and insert pestle.
C. Grind the sample 1-2 minutes (in subdued light) and wash
pestle and grinding tube with 5 ml of 90 percent acetone into a 15 ml
screw cap centrifuge tube. Centrifuge (2000 xfor 1-5 min., allow
to stand in the dark for 1-2 hrs to ensure the complete removal of all
extractable pigments.
D. Measure fluorescence as outlined in step B direct reading
procedure. If phaeophytin is to be measured, acidify with 2 drops
l.N HC1 and reread fluorescence.
E. Fluorescence values can be recorded as relative chlorophyll
values or as chlorophyll a^ (milligrams per cubic meter) as calculated
from the equation:
_3
chlorophyll a^ (mg m )
(Fo/Fa pn
* ' rna v'
(kx) (Fo - Fa)
liters filtered
3 (F°/FamaJ- 1 (kx) WF,nx
-------
61
where: Fo = fluorescence before acidification
Fa = fluorescence after acidification
Fo/Fa = maximum acid factor which can be expected in
the absence of phaeophytin
kx = calibration constant for a specific sensitivity
scale.
It should be noted that kx, Fo/Fa_, and acid ratios are functions of
max
the combination of photomultiplier and color filters.
References.
Lorenzen, C. J. 1966. A method for the continuous measurement of in
vivo chlorophyll concentration. Deep-Sea Res. 13,223-227.
Strickland, J. D. H. and T. R. Parsons. 1965. A manual of sea water
analysis. Fisheries Research Board of Canada, Bulletin No. 125,
2nd Revised Edition, 203 p.
Yentsch, C. S. and D. W, Menzel. 1963. A method for the determination
of phytoplankton chlorophyll and phaeophytin by fluorescence. Deep-
-------
APPENDIX 11
11.2 Reference curves - The results of an assay where cell counts
were used as the biomass indicator are presented in Table 1. Triplicate
sets of flasks were incubated for the control (lake water sample), the
control with a phosphorus addition, and the control with a nitrogen
addition. Figure 1 shows the average cell counts obtained in the same
triplicate sets of flasks plotted against time (days).
11.3 Maximum specific growth rate
11.31 Identification of growth rate limiting nutrients by
single nutrient spikes. The data shown in Table 1 were obtained by single
nutrient spikes of nitrogen and phosphorus to a lake water sample. The
daily specific growth rates (y) for each flask of each triplicate set of
flasks were calculated from the equation in Section 9.41 and are shown in
Table 2. The largest daily specific growth rate for each flask is its
y , and these were averaged to obtain ym3v for each set of triplicate
iTIaX max
flasks (Table 3). The phosphorus-spiked set of flasks had a u of
Ilia X
1.33 + 0.12, as opposed to a y of 0.91 +0.16 in the nitrogen-spiked
" it la a
flasks and of 0.71 +0.14 in the control flasks. Phosphorus, therefore,
can be identified as the growth rate limiting nutrient.
11.4 Maximum standing crop - The maximum standing crop can be
identified from the data contained in Table 4 where the daily cell
counts (Table 1) were averaged. The daily average cell count in the
control, for example, increased steadily for the first four days. After
that the daily increase was less than five percent and the maximum
3
-------
TABLE 1
TYPICAL REPORT OF ASSAY RESULTS
Time
C E L
L S PER
ML
(Days)
C 0 N T R 0
L
CONTROL + 0.05 mg P/l
C 0 N T
R 0 L + o.:
37 mg N/1
1
2
3
1
2
3
1
2
3
0
l.OxlO3
l.OxlO3
l.OxlO3
l.OxlO3
l.OxlO3
l.OxlO3
1 .OxlO3
l.OxlO3
l.OxlO3
1
2.lxlO3
1.7xl03
1.6xl03
3.3xl03
2.8xl03
3.1xl03
2.2xl03
1.9xl03
2.3xl03
2
4.1xl03
3.6xl03
3.0xl03
9.3xl03
8.1xl03
8.2xl03
5.9xl03
4.7xl03
4.4xl03
3
5.5xl03
4.4xl03
4.8xl03
3.4xl04
3.3x104
3.0xl04
6.2xl03
5.1xl03
6.1xlO3
4
7.8xl03
8.1xl03
6.6xl03
l.lxlO5
l.lxlO5
l.lxlO5
6.5xl03
CO
o
r>
X
00
7.1xl03
5
7.9xl03
8.2xl03
6.7xl03
1.3xl05
1.5xl05
1.4x105
7.6xl03
8.5xl03
8.3xl03
6
8.2xl03
8.2xl03
6.8xl03
1.5xl05
1.6xl05
1.6xl05
9.OxlO3
9.7xl03
9.6xl03
7
8.4xl03
7.2xl03
6.9xl03
1.6xl05
1.6xl05
1.6xl05
9.OxlO3
9.9xl03
-------
oo Control
^ & Control+.05mg/1 P
~ ~ Control+.37mg/1 N
' ' 1 1 1 1 '
I 2 3 4 5 6 7
TIME IN DAYS
Figure 1. Effect of the addition of nitrogen and phosphorus to a
-------
66
TABLE 2
TYPICAL REPORT OF GROWTH RATE CALCULATIONS
Time
(Days)
SPECIFIC GROWTH RATES (y) -
Day"1
CONTROL
CONTROL + 0.05 mg P/l
CONTROL +0.37
mg N/1
1
2
3
1
2
3
1
2
3
0
7
0.74
0.53
0.47
1.19
1.03
1.13
0.79
0.64
0.83
1
0.67
0.75
0.63
1.04
1.06
0.97
0.98
0.91
0.65
2
0
0.29
0.20
0.47
1.29
1.40
1.30
0.05
0.08
0.33
O
0.35
0.61
0.32
1.17
1.20
1.30
0.05
0.42
0.15
4
C
0.01
0.01
0.02
0.17
0.31
0.24
0.16
0.09
0.16
0
0.04
0.00
0.01
0.14
0.06
0.13
0.17
0.13
0.15
6
0.02
0.00
0.01
0.06
0.00
0.00
0.00
0.02
0.02
7
-------
67
TABLE 3
DETERMINATION OF MAXIMUM SPECIFIC GROWTH RATES
Flask
Number
Maximum Specific Growth Rate (y )
max
Control
Control + 0.05 mg P/l
Control + 0.37 mg N/1
1
0.74
1.29
0.98
2
0.75
1.40
0.91
3
0.63
1.30
0.83
Ave '"max'
0.71 ±0.14
1.33 ±0.12
-------
68
11.41 Identification of growth limiting nutrients - The
data in Table 4 can be used to identify the growth limiting nutrient.
The maximum standing crop for the control (7.5 x 10 cells/ml) is
compared to that for the nitrogen-spiked flasks (9.43 x 10 cells/ml)
and the phosphorus-spiked flasks (1.57 x 105 cells/ml). Since the
phosphorus added to the lake water resulted in a significant increase
in the maximum standing crop, it can be identified as the growth limiting
nutrient.
11.5 Confidence intervals - Confidence intervals are based upon
the standard deviation (a).
a = confidence interval of 66.6 percent
2a = confidence interval of 95.0 percent
3a = confidence interval of 99.0 percent
Example of calculation - Taking data (cell counts) from Table
1 for the triplicate set of control flasks for day one, the constants
in the equation are as follows:
x1 » 2.1 x 103 x^ = 4.41 x 103
x2 = 1.7 x 103 x22 = 2.89 x 103
-------
69
TABLE 4
CALCULATION OF MAXIMUM STANDING CROP
Time
(Days)
CONTROL
CONTROL + 0.05 mg P/l
CONTROL + 0.37 mg N/1
Ave. cells/ml 2a
Ave. cells/ml 2a
Ave. eel Is/ml 2a
0
l.OOxlO3 ±0.00xl03
l.OOxlO3 ±0.00xl03
l.OOxlO3 ±0.00xl03
1
1.80xl03 ±0.52xl03
3.07xl03 ±0.50xl03
2.13xl03 ±0.42xl03
2
3.57xl03 ±1.lOxlO3
8.53xl03 ±1.33xl03
5.00xl03 ±1.58xl03
3
4.90xl03 ±1.12x103
3.23xl04 ±0.42xl04
5.80xl03 ±1.22xl03
4
7.50x103 ±1.58xl03
l.lOxlO5 ±0.00xl05
7.13xl03 ±1.30xl03
5
7.60xl03 ±1.58xl03
1,40xl05 ±0.20xl05
8.13xl03 ±0.94x1O3
6
7.73xl03 ±1.60xl03
1.57xl05 ±0.12x1O5
9.43xl03 ±0.76x1O3
7
7.50xl03 ±1.58xl03
1.60xl05 ±0.00xl05
9.57xl03 ±0.98xlOS
2a = Standard Deviation at 95 percent confidence interval.
Maximum Standing Crop
Control = 7.50 x 10: ±1.58 x 103 (Day 4)
Control + 0.05 mg P/l = 1.57 x 105 ±0.12 x 105 (Day 6)
-------
70
Therefore: £x2 = 9.86 x 103
(2x)2 = (5.4 x 103)2 = 29.16 x 103
^Ex|2 = 29.16 x TO3 = g 72 x 103
n j
0 = ± J 9-86 x 1q32~ 9-72 x 103 = ± /0.07 = ±0.26
2a = ±0.52
3a = ± 0.78
Following is an example of determining the required number of
repli cates:
Consider the design of an experiment to compare two media,
one of known strength which will produce a maximum standing crop of about
ni| = 2 x 10^ cells/ml and another medium expected to produce a greater
standing crop. The "null hypothesis," which one expects to disprove,
is that m2 rnl' 1,e*' t'iat t'ie un' m-|, i.e., that the unknown medium
produces a greater standing crop than the known medium.
How many replicate pairs of flasks should be used? The answer
can be found by first answering the following five questions and then
-------
71
Question 1: "What significance level, a, should be used?"
For this example we shall use the significance level a = 0.05, i.e.,
if the two media are the same strength (m^ = n^) there will be one
chance in twenty that the experiment will result in the erroneous
conclusion that the known medium is weaker (m.| < m^).
Question 2: "What is the smallest difference, 6 = ir^ -
which must be detected?" The known medium will produce a standing crop
of about m-j = 2 x 10 cells/ml. Suppose the other medium must produce
g
a 10 percent greater crop (n^ = 2.2 x 10 cells/ml) to be "significantly"
stronger, i.e., the smallest difference which must be detected is about
6 = m2 - m-j = 2.0 x 106 - 2.2 x 106 = 2 x 105 cells/ml.
Question 3: "With what probability must a difference of
<50 (= 2 x 105 cells/ml) be detected by the experiment?" Suppose it is
desired to have a probability of detection of 0.90, i.e., if the true
5
difference in the standing crops of the media is 2 x 10 cells/ml there
is a 90 percent chance the experiment will detect the difference (lead
to the conclusion that the known medium is weaker). Conversely, there
is a 10 percent chance the experiment will fail to detect a difference
5
of 2 x 10 cells/ml. Denote the probability of detection as 1 - 3 = 0.90.
Question 4: "What is the standard deviation, s, of an
individual observation?" (Note that this is not the same as the standard
error of a mean of several observations.) There would probably be some
information about the standard deviation from a prior experiment with
the "known" medium. For this example assume that previous experience
indicates a standard deviation of approximately 2.7 x 10 cells/ml.
Question 5: "Does the alternative hypothesis specify a 'one-
-------
72
f m^)?" In this example it is assumed the findings will be significant
only if the unknown medium produces a greater standing crop than the
known medium; thus the alternative hypothesis specifies a one-tail
alternative (6 > 0, m2 > m-j). Therefore, a one-tail test will be used.
(A two tail alternative would require a two-tail test.)
In summary, the answers to the questions above have provided
the following values:
(1) a = 0.05 = significance level
(2) 6q = 2 x 105 cells/ml = smallest "significant" difference
(3) 1 - 0 = 0.90 = probability of detecting smallest signifi-
cant difference
(4) s = 2.7 x 10 cells/ml = standard deviation
(5) Alternative hypothesis specifies a one-tai1 test.
We can now compute the value of "d" and find the required
number of replicates from the table:
d = 0.7071 6q/s
= (0.7071) x (2 x 105)/(2.7 x 105)
= 0.524
Entering the One-Tail Test Tables with these values we find the number
of replicates should be between 54 + 1 (corresponding to d = 0.4) and
24 + 1 (corresponding to d = 0.6)*. One should use quadratic interpolation
in the table, but linear interpolation produces an approximate result:
36 replicate pairs. Note that only 10 replicate pairs would have the
desired probability of detecting the difference if d = 1.0, i.e., if
*Note that the tabled value is not the number of replicates; one must
add 1 to the tabled values in the a = .05 table and 2 to the tabled
-------
73
TABLE 5
AID IN COMPUTING SAMPLE SIZES REQUIRED TO DETECT
PRESCRIBED DIFFERENCES BETWEEN AVERAGES
Notation:
a Significance level of the test
60 Smallest detectable or significant difference
1-$ Probability of declaring
-------
74
TABLE 5 (continued)
TWO TAIL TEST TABLES:
For a = .01 add 2 to the tabled value to get the number of pairs
(replicates); for a = .05 add 1 to the tabled value to get the number
of pairs.
a = .01
1-3
.50
.60
.70
o
00
.90
.95
.99
.1
664
801
962
1168
1488
1782
2404
.2
166
201
241
292
372
446
601
.4
42
51
61
73
93
112
151
.6
19
23
27
33
42
50
67
.8
11
13
16
19
24
28
38
1.0
7
9
10
12
15
18
25
1.2
5
6
7
9
11
13
17
1.4
4
5
5
6
8
10
13
1.6
3
4
4
5
6
7
10
1.8
3
3
3
4
5
6
8
2.0
2
3
3
3
4
5
7
3.0
1
1
2
2
2
2
3
If we must estimate a from our sample and use Student's t, then we should
add 4 to the tabulated values to obtain the approximate required sample size.
(If we are comparing two product averages, add 2 to the tabulated values, to
obtain the required size of each sample. For this case, we must have a/\ = ag.)
a = .05
1-3
d
.50
.60
.70
.80
.90
.95
.99
.1
385
490
618
785
1051
1300
1838
.2
97
123
155
197
283
325
460
.4
25
31
39
50
66
82
115
.6
11
14
18
22
30
37
52
.8
7
8
10
13
17
21
29
1.0
4
5
7
8
11
13
19
1.2
3
4
5
6
8
10
13
1.4
2
3
4
5
6
7
10
1.6
2
2
3
4
5
6
8
1.8
2
2
2
3
4
5
6
2.0
1
2
2
2
3
4
5
3.0
1
1
1
1
2
2
3
2 2
u = (z^^+z^^) / d , where zx denotes the cumulative distribution
function of the standard normal (0,1) distribution.
Source: Experimental Statistics, by Mary G. Natrella, National Bureau of
Standards Handbook 91, U.S. Government Printing Office, Washington, D.C.
-------
6q = Q = ^^q27q7t[ 1q5) = 3-81 x 1q5- That is, 10 replicate
pairs would have a 90 percent chance of detecting a difference of 3.8 x
10 cells, a 19 percent increase in standing crop, whereas 36 replicates
are required to ensure a 90 percent chance of detecting a 10 percent
increase in standing crop. These figures assume the validity of the
estimate of the standard deviation.
11.6 Rejection of outliers - An "outlier" among replicate observations
is one whose deviation from the mean is far greater than the rest in
absolute value and perhaps lies three or four standard deviations or
further from the mean. The outlier is a peculiarity and indicates a data
point which is not at all typical of the rest of the data. It follows
that an outlier should be submitted to particularly careful examination
to see if the reason for its peculiarity can be determined.
Rules have been proposed for rejecting outliers, i.e., for
deciding to remove the observation(s) from the data, after which the
data are reanalyzed with these observations. Automatic rejection of
outliers is not always a wise procedure. Sometimes the outlier is
providing information which other data points cannot due to the fact that
it arises from an unusual combination of circumstances which may be of
vital interest and requires further investigation rather than rejection.
As a general rule outliers should not be rejected out of hand unless they
can be traced to causes such as errors in recording observations or in
setting up apparatus. Otherwise careful investigation is in order. (The
above was adopted from section 3.8 of Applied Regression Analysis by
-------
76
The following test may be applied for rejecting outliers:
1. Rank order the data in the group containing the outliers
(all observations in the group are supposedly treated alike):
x] 1 x2 xn
2. Compute the appropriate criterion:
If x, is the outlier c = ^ ~ ^
1 xn - X!
If x is the outlier c = Xn~ Xn~T
n xn - X!
3. If c exceeds the critical value opposite "n" in the following
table, reject the outlier.
n
Critical Values
a = 0.05
a = 0.01
3
0.941
0.988
4
0.765
0.889
5
0.642
0.780
6
0.560
0.698
7
0.507
0.637
Example - Suppose the following replicate observations were
made: 9.8, 4.7, 8.4, 8.0, 8.4, and 7.9. The value 4.7 is suspected to
be an outlier. Rank order the data.
x, x2 x3 x4 x5 x6 n = 6
4.7 7.9 8.0 8.4 8.4 9.8
The criterion is as follows:
x2 - xi 79 _ 4i7 _ 3.2 _ n
c ~ x - x, " 9.8 - 4.7 " 5.1 "
-------
Since n = 6, this value is significant at the a = 0.05 level (0.63 > 0.560),
but not at the a = 0.01 level (0.63 < 0.698).
The experimenter who is willing to discard five percent of
all his good data would discard the observation 4.7 as an outlier. The
experimenter who is willing to discard only one percent of all his good
data would keep the observation unless he can determine an experimental
reason for rejecting it.
4. If there are two suspected outliers (say x^ and or
x-j and Xg, the test may be repeated; apply it to the "worst" outlier
first.
5. Note that the regular use of this procedure will result
in discarding five percent (if a = 0.05) or one percent (if a = 0.01)
of all one's good (valid) observations.
11.7 Evaluation of assay results - A computer program in BASIC,
which can analyze a 2x2x2x2 FACTORIAL experiment, is presented in Table 6.
Typical results of analysis of variance for such a growth rate experiment
are shown in Tables 7, 8, and 9 for an influent to and effluent from a
ground percolation process. In this example none of the nutrients or
groups of nutrients limit the maximum specific growth rate (ymax) of
the test algae in the influent (Table 8). In the effluent (Table 9),
both Factor 3 (P + N) and Factor 4 (macronutrients) were significant
at the 95 and 99 percent confidence levels respectively.
The evaluation of significance, however, should always include
a comparison of absolute effects. In this case a comparison of Mean 1
(with spiking) with Mean 2 (without spiking) shows that, although the
spiking with macronutrients was statistically significant at a confidence
level greater than 99 percent, the absolute increase in pmax due to spiking
-------
78
TABLE 6
PROGRAM FOR ANALYSIS OF 2x2x2x2
FACTORIAL EXPERIMENTS
Prepared in BASIC for a HP2116B Computer
LIST
1 REM ANALYSIS OF VARIANCE PROGRAM K. JUSTICE. SCHOOL OF
5 REM ENGINEERING UNIV. OF CALIF. IRVINE,CALIF. 92664
10 REM THIS PROGRAM WILL PERFORM AN ANALYSIS OF VARIANCE AND
15 REM COMPUTE F-RATIOS FOR A 2X2X2X2 (2*4) FACTORIAL EXPERIMENT
20 REM WITH TWO REPLICATES. TO USE, TYPE:
25 REM "200 DATA X(1 ,1) ,X(1 ,2) ,X(2,1) ,X(2,2) X{M,N)"
30 REM WHERE M IS THE NUMBER OF SAMPLES AND N IS THE NUMBER
35 REM OF REPLICATES.
40 REM USE AS MANY 'DATA STATEMENTS' AS NECESSARY, NUMBERING
45 REM THEM SUCCESSIVELY.
100 DATA 16,2
300 DIM U[17,3],T[17],R[17]
305 READ M,N
309 FOR 1=0 TO 16
310 LET T[I+1]=R[H1]=0
312 NEXT I
315 LET G=R2=R=W=0
330 FOR 1=1 TO M
335 FOR J=1 TO N
340 READ U[I+1,J+1]
345 LET T[I+1]=T[I+1]+U[I+1,J+1]
350 LET R=R+U[I+1 ,J+1]+2
355 NEXT J
360 LET G=G+T[I+1]
365 LET W=W+T[I+1]+2
370 NEXT I
375 LET A= 4.54 : * SIGNIFICANT AT 95% LEVEL"
F > 8.68 : ** SIGNIFICANT AT 99% LEVEL"
-------
UN:
TOR
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
TABLE 7
COMPUTER PRINTOUT CODE
Refers to the elements added as shown in code above.
Each factor line in printout refers to a statistical
comparison between those flasks which had the elements
added versus those flasks which did not have the same
elements added.
Sign indicates direction of effect of addition. A
negative sign means that addition of that particular
combination of factors resulted in a decrease.
* Indicates significance at the 95 percent level.
**Indicates significance at the 99 percent level.
1 Indicates mean of flasks with elements added.
2 Indicates mean of flasks without elements added.
ELEMENTS ADDED (1/3 MEDIUM CONCENTRATION)
(Trace)
(Fe + Mn)
(P + N)
(Macro)
(Trace)
(Trace)
(Trace)
(Fe + Mn
(Fe + Mn
(P + N)
(Trace)
(Trace)
(Fe + Mn
(Trace)
(Trace)
Fe + Mn)
P + N)
Macro)
(P + N)
(Macro)
Macro)
Fe + Mn)
Fe + Mn)
(P + N)
P +
Fe
(P + N)
(Macro)
(Macro)
N) (Macro)
Mn) (P + N)
-------
81
TABLE 8
INFLUENT TO PERCOLATION PROCESS
Maximum Growth Rate (Day-^)
READY
RUN
ANALYS
16
SAMPLES
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
IS OF VARIANCE
SAMPLES OF SIZE 2
SAMPLE TOTAL SAMPLE MEAN
2.907
1.4535
3.072
1.536
2.956
1.478
3.001
1.5005
3.087
1.5435
3,17
1.585
2.941
1.4705
2.988
1.494
2.959
1.4795
3.041
1.5205
3.135
1.5675
3.022
1.511
3.062
1.531
2.905
1.4525
3.014
1.507
2.86
1.43
MEAN SQUARE (BETWEEN SAMPLES) = 3.65499E-03
MEAN SQUARE (WITHIN SAMPLES) = 7.08389E-03
MEAN SQUARE (REPLICATES) = 1.08795E-02
ERROR = 6.83085E-03
FACTORIAL ANALYSIS WITH FOUR TREATMENTS AT TWO LEVELS
FACTOR
COMPARISON
F-RATIO
MEAN 1
1
2.00033E-03
1.83054E-05
1.50381
2
.286002
.374207
1.51269
3
6.60033E-02
1.99300E-02
1.50581
4
.124002
7.03449E-02
1.50763
5
-.348
.554031
1.49288
6
-.36
.592899
1.4925
7
-.681999
2.12786
1.48244
8
-.556001
1.41425
1.48638
9
.413999
.784104
1.51669
10
-.566001
1.46558
1.48606
11
-.282001
.363812
1.49494
12
3.59998E-02
5.92891E-03
1.50488
13
-5.60002E-02
1.43468E-02
1.502
14
.2
.182994
1.51
15
.114
5.94544E-02
1.50731
F > 4.54 : * SIGNIFICANT AT 95% LEVEL
F > 8.68 : ** SIGNIFICANT AT 99% LEVEL
MEAN 2
1.50369
1.49481
1.50169
1.49988
1.51463
1.515
1.52506
1.52113
1.49081
1.52144
1.51256
1.50263
1.5055
1.4975
1.50019
-------
TABLE 9
EFFLUENT FROM PERCOLATION PROCESS
Maximum Growth Rate (Day-^)
ANALYSIS OF VARIANCE
16 SAMPLES OF SIZE 2
SAMPLES
SAMPLE TOTAL
SAMPLE MEAN
1
2
3.424
1.712
3
4
3.509
1.7545
5
6
3.425
1.7125
7
8
3.625
1.8125
9
10
3.409
1.7045
11
12
3.527
1.7635
13
14
3.467
1.7335
15
16
3.352
1.676
17
18
3.181
1.5905
19
20
3.339
1.6695
21
22
3.199
1.5995
23
24
3.409
1.7045
25
26
3.222
1.611
27
28
3.06
1.53
29
30
3.2
1.6
31
32
3.153
1.5765
MEAN SQUARE (BETWEEN SAMPLES) = 1.25570E-02
MEAN SQUARE (WITHIN SAMPLES) = 3.24726E-03
MEAN SQUARE (REPLICATES) = 2.07520E-03
ERROR = 3.32540E-03
FACTORIAL ANALYSIS WITH FOUR TREATMENTS AT TWO LEVELS
FACTOR
COMPARISON
F-RAT10
MEAN 1
MEAN 2
1
-.447
1.87768
1.65794
1.68588
2
-.159
.237576
1.66694
1.67688
3
.721003
* 4.88517
1.69444
1.64938
4
1.975
** 36.6556
1.73363
1.61019
5
4.89998E-02
2.25629E-
¦02 1.67344
1.67038
6
-.859
* 6.93414
1.64506
1.69875
7
-.129001
.156383
1.66788
1.67594
8
-.250998
.592037
1.66406
1.67975
9
.159
.237576
1.67688
1.66694
10
-.264999
.659927
1.66363
1.68019
11
.284999
.763295
1.68081
1.663
12
-.285
.763301
1.663
1.68081
13
-.217
.442513
1.66513
1.67869
14
.295
.817806
1.68113
1.66269
15
.411
1.58741
1.68475
1.65906
F >
4.54 : * SIGNIFICANT AT 95%
LEVEL
F > 8.68 : ** SIGNIFICANT AT 99% LEVEL
------- |