Environmental Protection Agency
November 1981

0 \\
I. Introduction	1
II. The Regulatory Framework and Applicable
Standards £or Reviewing Requests for
q—	Reconsideration of a Prior Cancel-
lation Order	2
^ III. The Cancellation of 1080 and Its
Subsequent Regulatory History	5
^ IV. Summary of Livestock Death Losses	13
V. Efficacy of 1080 and Alternative
Methods of Reducing Predation	45
VI. Hazards to Non-Target Wildlife and
Humans	88
References	125

) The use of the highly toxic pesticide sodium fluoroacetate,
commonly called "1080/" has long been a controversial issueTj
Since its introduction in the United States in 1944/ 1080 has
been used to kill coyotes suspected of preying on livestock.
Livestock ranchers have consistently praised its efficacy and
safety, while environmentalists have claimed that 1080 has caused
the death of countless non-target wild animals, from the skunk
to protected species like the bald eagle and California condor.
Despite EPA's decision in 1972 to suspend and cancel the federal
registration of any pesticide containing 1080 for use as a preda-
cide, the controversy has persisted. ^Livestock producers and
others have[constantly Iprged EPA to reconsider its position on
chemical toxicants, particularly compound 1080.. In order to
assist the Agency in deciding whether the cancellation order
should be reconsidered, EPA in late July 1981 conducted informal
public hearii in Denver, Colorado and Washington, D.C. to
solicit the views of interested persons concerning these issues.

- 2 -
II. The Regulatory FrameworJc and Applicable Standards for
Reviewing Requests for Reconsideration of a Prior
Cancellation Order
Under the Federal Insecticide, Fungicide, and Rodenticide Act
(FIFRA) 7 U.S.C. 136 et seq., EPA has the congressionally mandated
responsibility for ensuring that pesticides marketed in the United
States do not pose "unreasonable adverse effects on the environ-
ment, " a term which is defined as "any unreasonable risk to man
or the environment taking into account the economic, social, and
environmental costs and benefits of the use of the pesticide."
FIFRA SS2(bb), 3(c)(5), 6(b). This responsibility is exercised
primarily by Agency decisions on whether to register (license)
pesticide products. Toxicants intended to control predators are
pesticides under FIFRA, and therefore are subject to EPA registra-
Under current law, if it appears to the Administrator that a
pesticide no longer meets registration requirements and appears
to cause unreasonable adverse effects on the environment, the
statute provides that the Administrator, after affording an
opportunity for hearing, may cancel or suspend the pesticide's
registration. FIFRA §6(b), (d). The burden of proof of a pesti-
cide's registrability remains on proponents of registration at
all times. (See Environmental Defense Fund v. Environmental Pro-
tection Agency, 548 F. 2d 998 (D.C. Cir. 1976), cert, denied
431 U.S. 925).
In Subpart 0 of the Agency's Rules of Practice governing hear-
ings under FIFRA, EPA has established regulations for reviewing
applications for registration or for emergency use exemptions
for pesticides which have been previously cancelled or suspended.
(See 40 CFR $164,130 et seq). These regulations provide that any
application for registration under FIFRA §3 or request for an
emergency exemption under FIFRA $18 to allow use of a pesticide
whose registration has been finally suspended or cancelled shall
be considered a petition to reconsider the prior cancellation

or suspension order. (See 40 CFR §164.130.)
Under the Subpart D regulations, the Administrator shall
determine that reconsideration of a prior cancellation or suspen-
sion order is warranted only if the Administrator finds that:
1)	the applicant has presented substantial new evidence
which may materially affect the prior cancellation or
suspension order and which was not available to the
Administrator at the time he made his final cancellation
or suspension decision and
2)	such evidence could not, through the exercise of due
diligence, have been discovered by the parties to the
cancellation or suspension proceeding prior to the
issuance of the final order.
(See 40 CFR §164.131(a).)
The regulations further provide the Administrator shall pub-
lish in the Federal Register her determination as to whether
reconsideration of the prior order is warranted and the basis for
the determination. If reconsideration is warranted, the notice
shall announce that a formal adjudicatory hearing will be held to
examine the evidence and to determine whether to modify or re-
verse the prior order. (See 40 CFR §164.131(c)). During the hear-
ing the applicants) has the burden of proof. (See 40 CFR §164.
132(a)). Under the regulations, the issues in the hearing con-
sist of whether substantial new evidence exists and, if so,
whether such substantial new evidence requires reversal or modi-
fication of the suspension order. (See 40 CFR §164 .132(c)). At
the end of the hearing, the presiding officer will mate findings
of fact and regulatory conclusions. This preliminary decision
may be appealed to the Administrator for final Agency decision.
Thus, two procedural stages normally occur before a cancel-
lation or suspension decision can be reversed or modified:

1) the presentation of substantial new evidence which may materi-
ally affect the previous order and 2) a notice and opportunity
for an adjudicatory hearing concerning whether the existing
cancellation or suspension order should be reversed or modified.*
Finally, the entry of an order modifying or reversing the
1972 cancellation decision concerning predacide use of 1080
would not constitute registration or an approval of an emergency
exemption. Rather, the order would only specify the conditions
(if any) under which 1080 could be used to kill predators. EPA
would refuse to register any product that did not satisfy those
conditions. An applicant, however, would still need to fulfill
additional requirements in order to obtain a section 18 emergency
exemption** or a section 3 registration (see 40 CFR Sections
162.18-1 through -5).
* The regulations also provide that the Administrator may
dispense with the requirement of convening a hearing when she
determines (49 CFR §164.133):
(1)	That the application presents a situation involving
need to use the pesticide to prevent an unacceptable
risk: (i) To human health, or (ii) to fish or wild-
life populations when such use would not pose a hu-
man health hazard; and
(2)	That there is no other feasible solution to such
risk; and
(3)	That the time available to avert the risk to human
health or fish and wildlife is insufficient to per-
mit convening a hearing as required by §164.131;
(4)	That the public interest requires the granting of
the requested use as soon as possible.
** Generally, a request for an emergency exemption from FIFRA's
requirements must show that significant pest control problems
are occurring or are about to occur, and that there is no regis-
tered pesticide or non-pesticide method that can effectively
control the problem. The request must also provide information
on the nature of the control program, the risks associated with
the,program and the anticipated benefits. (See 40 CFR Part 166.)

- 5 -
III. The Cancellation of 1080 and Its Subsequent Regulatory
A. The Cancellation of pesticides Registered for
Predator Control
In the 1960's wildlife protection groups began to express
concern that, because of its acute toxicity, 1080 was causing the
death of wildlife other than coyotes. This public concern led in
1972 to a decision by EPA to cancel and suspend all federal
registrations for the use of 1080 products—as well as products
containing three other toxicants, sodium cyanide, strychnine, and
thallium--as predacides. Earlier in the same year, President
Nixon issued Executive Order 11643 which prohibited the use of
all chemical toxicants including 1080 on federal lands or in
federally administered animal damage control programs, except in
an emergency.*
* Under the Executive Order, an emergency requiring the use of
a chemical toxicant may be declared only by the head of an agency
having jurisdiction over the federal land where the toxicant
would be used, or having responsibility for the predator control
program if the use would be by federal employees on private land.
Prior tp declaring an emergency, the agency head must consult
with the Secretaries of Interior, Agriculture, and Health and
Human Services and with the Administrator of EPA. In addition,
the agency head must make a written finding that the emergency
cannot be dealt with by means which do not involve use of a
toxicant and that such use is necessary to protect human health
or safety, to preserve threatened or endangered species, or to
prevent substantial and irretrievable damage to nationally
significant natural resources.
The Executive Order has been amended twice, but the amend-
ments do not affect the provisions concerning emergency excep-
tions to the general prohibition against federal use of chemical

- 6 -
SPA's 1972 cancellation and suspension decision relied heavily
on information from three sources, "The Cain Report," * the Natural
Resources Defense Council petition to ban 1080, and "The Leopold
Report." "The Leopold Report," officially titled "Predator and
Rodent Control in the United States," was written in 1964 by a
special advisory committee appointed by the Secretary of Interior
to study wildlife management. Chaired by Dr. A. S. Leopold, the
group prepared a brief report recommending sweeping changes in
the federal predator control program. In its discussion of the
existing program, "The Leopold Report" (p. 10) noted that 1080
was often misused and that considerable damage to other forms of
wildlife could occur as a result. "The Leopold Report" also stat-
ed, however, that "when properly applied . . . 1080 [does] as
effective and humane job of controlling coyotes and [has] very
little damaging effects on other wildlife." Id.
The petition submitted in 1971 by the Natural Resources De-
fense Council and six other environmental groups strongly opposed
even the legal use of 1080, charging that use of 1080 and other
chemical toxicants had destroyed hundreds of thousands of "non-
target" animals including members of rare and endangered species.
The petition contained numerous references to articles and stud-
ies to substantiate its contentions.
Finally, EPA relied on "The Cain Report," which was pub-
lished in 1972. This report, the most comprehensive of the
three sources, surveyed a large body of data on predation loss
rates, predator control methods, wildlife toxicology, and a
number of other subjects. "The Cain Report" (pp. 5-6) strongly
recommended that any use of toxic chemicals for predator control
be prohibited.
* The report is officially known as "The Report to the Council
on Environmental Quality and the Department of the Interior by
the Advisory Committee on Predator Control." The report is
commonly called "The Cain Report," and the Advisory Committee on
Predator Control is called the Cain Committee after the chairman,
Dr. Stanley A. Cain.

- 7 -
EPA's cancellation decision contained the following findings
of fact:
-	1080 is highly toxic to all species. The dangerous
dose for man is 0.5-2.0 tag/kg. The chemical acts
rapidly on the central nervous and cardiovascular
systems with cardiac effects. Effect is usually
too quick to permit treatment, and antidotes are
relatively valueless.
-	According to one authority, prior to 1963, there
were 13 proven fatal cases, five suspected deaths,
and six non-fatal cases of 1080 poisoning in man.
-	There is evidence that a certain number of non-
target animals are being adversely affected by 1080
products, particularly in the case of carrion-eating
birds and mammals, by secondary poisoning.
-	There is no reliable data as to the loss of sheep
that might occur, without a predator control program
using these poisons.
-	Effective non-chemical alternatives exist, including
denning, shooting, and trapping—methods that have
long been available and effective, though more costly
than poisons.
B. The Regulatory History of 1080 Since the 1972 Cancellation
Since cancellation of the predacide use of compound 1080,
numerous groups have tried to obtain authority to use 1080 to
kill coyotes preying on livestock. With the exception of three
experimental use permits issued in 1977, 1980, and 1981 for the
1080 toxic collar, SPA has not permitted the use of 1080 as a

predacide. This section briefly describes the various attempts
to obtain reconsideration of the 1972 cancellation decision.
1. Initial Efforts to Overturn or Modify the 1972 Decision
Host of the attempts to overturn or create exceptions to the
1972 ban on the use of 1030 as a predacide have been made by
state Departments of Agriculture. During the last nine years
they have submitted several applications for registration (under
FIFRA S3) or emergency exemptions (under FIF3A §13). The first
group of applications was submitted in 1973/ and a second, larger
group was submitted in 1977. EPA has denied all of these applica-
tions on the grounds that none of the applicants submitted sub-
stantial new evidence which would indicate that the 1972 decision
should be reconsidered. See 43 FR 14,100; April 4, 1978 (EPA
ruling that Wyoming and other state applicants had not provided
sufficient evidence to warrant holding a Subpart D adjudicatory
hearing on 1030)*
In addition to these attempts to use the administrative
process to modify or reverse the 1972 cancellation order, during
the summer of 1974 six western states sued EPA in federal district
court. The plaintiffs asked the court to rule that EPA's 1972
cancellation order was illegal because EPA had failed to prepare
an Environmental Impact Statement (EIS) and because EPA had not
conducted an administrative hearing prior to issuing the 1972
order. The district court issued a preliminary injunction
prohibiting EPA enforcement of the 1972 order. On appeal the
U.S. Court of Appeals for the Tenth Circuit held that EPA was not
required to prepare an EIS on the cancellation of 1030. The Tenth
Circuit also held that EPA was not required to hold an adjudicatory
hearing before allowing the cancellation and suspension order to
become final because no registrant or applicant had requested such
a hearing. Moreover, once that order had become final, the Agency
was not required to hold a hearing, even if requested to do so.

- 9 -
Finally/ the court of appeals noted that, under FIFBA, a person
adversely affected by the cancellation and suspension order (e.g./
user groups such as the plaintiffs) could have obtained judicial
review of that order/ but only by petitioning for review within
sixty days after the order was entered. Since no one had sought
judicial review within the time period allowed by statute/ the
soundness of the 1972 decision to ban 1080 could not be challenged
in court. (See Wyoming v. Hathaway, 525 P.2d 66 (10th Cir. 1975)
cert, denied 426 U.S. 906). Accordingly, the injunction entered
by the district court was vacated, and the suit was remanded for
further proceedings. In 1979 the proceeding before the district
court was dismissed.
2. Experimental Use Permits
In August 1975, EPA Administrator Russell Train established
a policy governing the issuance of experimental use permits for
previously cancelled pesticides. Because registration of such
compounds requires, as an initial step, the showing of substan-
tial new evidence. Administrator Train required that any field
testing with these compounds should hold some reasonable promise .
of producing evidence that might persuade the Administrator to
reconsider the prior cancellation order. EPA would not issue
permits for experimental programs which did not meet this stan-
dard .
Since then, EPA has received several requests for experiment-,
al use permits involving 1080. In December 1975, Wyoming asked
EPA for permission to use 1080 in a series of experimental field
studies. EPA made several suggestions about the design of the
experimental programs and indicated that if such changes were
made/ it would probably issue an experimental use permit. Wyoming
did not respond to these suggestions, and no experimental use
permit was issued.
Texas requested a permit the next year to use 1080 in single
lethal dose drop baits. EPA rejected the request, noting that
the experimental design was unlikely to produce any substantial
new information, especially concerning the hazards to non-target

— 10
species. EPA recommended that Texas perform some preliminary
studies with non-lethal drop baits to determine which kind o£
baits were accepted by predators and least attractive to non-
target species. Such studies might indicate that a particular
kind of bait was sufficiently selective for predators to justify
research using drop baits containing 1080. EPA rejected, for the
same reason, a similar request submitted by Montana in 1977.
Neither Texas nor Montana has submitted studies of the sort
suggested by EPA in an application for an experimental use permit
for single lethal dose drop baits.
In 1977, the U.S. Department of the Interior (USDI) requested,
and EPA issued, an experimental use permit for 1080 in a "toxic
collar." EPA concluded that this delivery mechanism might
significantly reduce the hazards to non-target species and thus,
that such research could produce substantial new evidence which
might warrant reconsideration of the 1972 cancellation decision.
EPA has renewed this permit each year since 1977, and the permit
is now scheduled to expire in November 1981. The results of this
testing constitute the largest part of new information on the
1080 toxic collar. In addition, EPA has approved an experimental
program for use of 1080 in a toxic collar by the Texas Agricultur-
al Experiment Station, a part of Texas A&M University. This
permit was issued in May 1980 and expired a year later. EPA has
also issued a permit to New Mexico to experiment with use of 1080
in a toxic collar. EPA expects information from New Mexico's
studies to be valuable because of the differences in terrain and
husbandry practices between Texas and New Mexico. This one year
permit expires in February 1982.
Finally, EPA has rejected two other requests for experimental
use permits for 1080 in vessels to be attached to livestock. The
first, made by the Texas Department of Agriculture, was denied
because at the time EPA had issued two other permits (to USDI
and the Texas Agricultural Experiment Station) which authorized
studies with 1080 toxic collars in Texas. EPA concluded that the
Texas Department of Agriculture was unlikely to develop any infor-
mation which would differ from that being generated under the

- 11 -
other two permits. The University of Wyoming also requested an
experimental permit for use of 1080 in a program to control
livestock predation. In the fall of 1980, the university asked
EPA to authorize experimental programs with 1080 in balloons
glued to the flanks of sheep and in single lethal dose drop
baits. EPA asked the university to provide more information to
support its request, but none has been submitted.
3. Recent Requests From the States and Livestock Groups
for Permission to Use 1080
EPA has recently received a variety of requests for regis-
tration or emergency exemptions for the use of 1080 as a predacide.
Three states, one federal agency, and two livestock associations
have asked EPA approval for 1080 use in the toxic collar, large
bait stations, and drop baits. These applications are summarized
in Table III-l and Table III-2.
The U.S. Department of the Interior has submitted an appli-
cation for conditional registration under FIFRA 53(c)(7) of the
toxic, collar.* The product for which registration is sought is
the collar currently being tested in experimental programs in
Texas and New Mexico. The collar would be placed on sheep and
other livestock in areas where coyote predation was expected.
The State of Wyoming has requested that EPA grant an emer-
gency exemption from the registration requirements of FIFRA or
alternatively a conditional registration for a 1080 product to
use against coyotes preying on livestock. In both instances,
Wyoming seeks to use 1080 in single lethal dose drop baits and
large meat bait stations. Wyoming has also issued a state regis-
tration, under FIFRA §24(c) , for 1080 use in single lethal dose
* FIFRA 53(c)(7) and the implementing regulations provide generally
that a person may obtain registration for a pesticide which is
intended for a use that is not authorized by any existing registration
for that pesticide, so long as the applicant submits data pertaining
to the safety of the new use and EPA finds that registration and
use of the product will not significantly increase the risk of
unreasonable adverse effects on the environment.

Table III-l. Applications for Section 18 Emergency exemptions for use of 10B0 as a Predacido,


Basis for
Montana Dept. of
Single lethal
3.6 mg/1080
Up to 10
"Livestock depredation must be
drop bait
feral dogs
baits per
documented by department agent

with 15 grams
before use is granted.
ID #: 81-MP-Q7

of ground or

rendered lard,

"Dates of use} Sept. 1, 1981-
Date Submitted*

tallow or ani-

August 31, 1982.

mal matter.

to control

"1080 will be used only by

A total of no

trained, certified government

more than 75

applicators under the

grams of 1080

supervision of the Montana

will be used.

Department Livestock.
Wyoming Dept. of
Single lethal
Drop baits-
"Use only by qualified and
drop bait.
a maximum
density of
certified applicators

meat baits
of 5mg of
10 baits
after losses are
ID # s 01-W5f-02

1080 per
per square


Date submittedt

red foxes)
maximum of

"Bait placement must be

1 pound.

approved by program supervisor

Meat baita-
Meat baits-

"Access roads arid trails must

1.6 grams of
one meat

be posted in English and

1080 per 100
bait per


pounds of
36 square


"No bait stations to be placed

maximun of

near residences, areas of

2 pounds.

intensive fanning or rec-
reational areas.
"Records pertaining to bait
location, monitoring, etc.
to be kept by Department
of Agriculture and certified
"Bait stations to be completely
destroyed at tije end of the
baiting season.
"Dates of use? Sept. 1, 1981-
April 30, 1982.

Table III-l. Applications for Section IB Bnergency Exemptions for use of 1080 as a Predacide.


Basis for
South Dakota
Single lethal
Drop baits-will
No more than
Increase in
Dept. of Ag.
drop bait,

be composed of
2 baits at

toxic collars

15grams of
one station
ID#: 82-SD-01

lard, tallow
and no more

or other
than 5
Date Submitted:

animal tissue;


baits will
located in
Would control

contain no
a section.
costs of

more than
(640 acres)
viable animal

3.6 mg. of

damage con-

No more than
trol program

10 collars

will be

placed on

sheep at

a given


"Use by certified applicators
"Baits/collars to be used only
after coyote kill is verified
"Warning signs will be posted
at points of entrance to
field3 where baits are used.
"Qualified ADC personnel will
select bait sites.
"Baits will be monitored at
least every 7 days.
"Baits and warning signs to be
removed when predator is
eliminated or within 30 days
of placanent.

Table III-2. Applications for Section 3 Registrations for use of 1080 as a Predacide.
Target Dose
Species Rate
Special Restrictions
USDI— Fish and
Wildlife Service
Pile Symbols
Dates 9-22-81
toxic collar coyotes
one collar
per animal
(sheep or
1.08% 1080
'restricted use pesticide
Wyoming Dept. not specified mammals single dose	"restricted use pesticide
of Agriculture	"government agency use only
File Symbols
Datej 7-28-81

drop baits and large meat bait stations. EPA has informed Wyo-
ming that this latter action is specifically prohibited by FIFRA,
and therefore it has no legal effect.*
The Montana Department of Livestock and the South Dakota
Department of Agriculture have also requested that EPA grant an
emergency exemption under FIFRA §18 for products containing 1080
for use against coyotes and feral dogs which prey on livestock.
Montana's product would consist of a single lethal dose of 1080
and would be applied only by state personnel. South Dakota pro-
poses to use two delivery mechanisms, the 1080 toxic collar and
single lethal dose baits, under South Dakota's proposal, only
certified applicators (individuals specially trained to use high-
ly toxic pesticides) or people under their direct supervision
could use the 1080 products.
The Montana Department of Agriculture has also notified EPA
that it plans to issue a special local needs registration for
use of 1080 as a predacide under §24(c) of FIFRA, but has also
stated that it will not allow use of 1080 under this registration
until EPA has approved either a federal registration or an emer-
gency exemption for this use. The State of Colorado has submitted
a similar request, asking EPA approval of a special local needs
registration for use of 1080 in large bait stations and single
lethal dose drop baits. EPA has notified both Montana and Colo-
rado that FIFRA 524(c) would prohibit the state from issuing the
special local needs registration described in their notices.
Finally, in the spring of 1981, EPA received a letter from
the National Woolgrowers Association and the National Cattlemen's
Association which claimed that there have been increasing losses
from predation due to the unavailability of chemical controls such
* Section 24(c) of FIFRA authorizes states to "provide registration
for additional uses of federally registered pesticides formulated
for distribution and use within that State to meet special local
needs . . . ." However, this section also provides that a state's
authority does not extend to any use "if registration for such
uses has . . . previously been denied, disapproved, or cancelled
by the Administrator." (See also EPA's "24(c) regulations" at 46
Fed. Reg. 2008 [January 7, 1981]; to be codified at 40 CFR
SS162.150 - .156.) Thus, Wyoming lacks authority to issue a regis-
tration for 1080 as a predacide.

as compound 1080. The two groups requested that EPA concur in
the declaration of an emergency resulting from coyote predation
on livestock in western states and that EPA allow emergency use
of 1080 to control coyotes and feral dogs which prey on sheep and
cattle. No formal application has been submitted.*
C. Public Hearings
Given the claims of ever increasing losses of sheep and
cattle due to predation and the strong interest of certain states
and livestock producers in the restoration of 1080 use, the Agency
concluded that it would be useful to hold public hearings under
PI PSA 521(b) to gather information—information which might be
used to support a request for reconsideration of the 1972 cancel-
lation order. Accordingly, CPA scheduled hearings in Denver,
Colorado, for July 28 and 29 and in Washington, D.C., for July
31, 1981.
The public hearings attracted over 80 witnesses, including
livestock ranchers, wildlife biologists, trappers, scientists,
environmentalists, state officials, representatives of livestock
associations, and representatives of animal welfare groups. The
witnesses presented a diversity of viewpoints; some opposed use
of 1080, while others favored its use in various forms. A panel
of five federal employees from EPA and the departments of Agricul-
ture and Interior received both written and oral testimony. At
the close of the hearings, the panel members prepared brief re-
ports identifying the witnesses who provided information which
appeared to be both new and potentially significant. Though
' This letter contains both a request for an emergency exemp-
tion under FIFRA §18 and a request that EPA concur with the
Department of the Interior and other federal agencies in de-
claring, under the terms of Executive Order 11643, that an emer-
gency exists which would justify using 1080 on public lands.
Since no one may legally use 1080 under the Executive Order until
EPA issues either a section 18 emergency exemption or a section
3 registration, EPA will not address the appropriateness of
revising the Executive Order at this time.

- 17 -
cross-examination of witnesses by members of the public was not
permitted, the panel members asked some of the witnesses to
explain or clarify their comments and to provide references to
support their comments.
S. Conclusion
The preceding review indicates that the Agency, in accordance
with its regulations, has consistently demanded that proponents of
1080 use present substantial new evidence that would justify
reconsideration of the 1972 cancellation order. In the past,
proponents of 1080 use have failed to present significant new
information. This review also indicates that EPA has issued
experimental use permits for experimental programs which promised
to generate relevant new data. In many cases, however, people
requesting experimental use permits have been unwilling or unable
to design satisfactory studies. Thus, only three experimental
use permits have been issued for 1080 use.
In 1964, "The Leopold Report" noted a lack of scientific data
in many areas relevant to the predator control controversy and
urged further research. Seven years later "The Cain Report" ex-
pressed the same concerns, calling for more research. .Now, nine
years after "The Cain Report," EPA has received some new data. As
a result o£ the studies performed under the experimental use
permits and some laboratory research on wildlife toxicology, EPA
now has more information on the environmental hazards of 1080.
EPA has also received more information concerning the livestock
loss rates due to predation in various states and localities.
This information is discussed more fully in the next three sec-
tions of this document.

The purpose of this section is to summarize and discuss the
data concerning livestock losses to predators that were the
basis o£ SPA's 1972 order cancelling the use of 1080 and the
data which have become available since 1972. There are several
research methods that have been used to gather, measure and
evaluate livestock losses to predators. The most common methods
are direct field studies and mail and interview surveys. Direct
field studies are intense, biological studies conducted over
small areas for the purpose of documenting causes of livestock
death. Mail and interview surveys use questionnaires to deter-
mine loss levels and causes of death over large areas (i.e.,
states or regions). Other information which is not gathered
using rigorous research methods but which provides useful in-
sights into overall losses of livestock to predatorsr includes
information provided by individual livestock producers, livestock
producer organizations, various state agencies and universities.
A significant amount of information of this type was presented
during three days of informal public hearings held in late July,
and it is summarized herein .along with the scientific literature
published since 1972.
As noted in the previous section, EPA based its 1972 decision
primarily on "The Gain Report". That report found only limited
scientific literature regarding livestock losses to predators.
Based on "The Cain Report" and other information, the Agency
1.	There are no reliable data as to the amount of
predator control achieved by use of poisons.
2.	There are no reliable data as to the loss of
sheep that might- occur without a predator con-
trol program using poisons or the real effect
of such losses on the general economic health
of the sheep industry.

3.	Predator losses may be only a minor part of
total losses.
4.	Several non-chemical alternatives exist though
they are more costly than poisons.
5.	The federal government has committed itself to
a research program for methods of controlling
predators other than with poisons.
37 Fed. Reg. 5718* 5720. Since the 1972 order, several research
studies and various surveys have been conducted which specifical-
ly address predation on livestock. Few of these studies (either
before 1972 or after 1972) were conducted under similar condi-
tions or over identical time periods. Due to variable conditions
from study to study, it is difficult to compare predation rates
over time.
A- The Cain Report
The Cain Commission discussed loss data from five sources:
Nielson and Curie (1970) - Utah only. Data collected
for fiscal 1968-69. Reported ewe losses of about
2 percent, lamb losses of about 40 per 1,000 ewes for
a total loss of about 61 ewes and lambs per 1,000
Owen Morris* Estimates - Utah only. Data assembled from
early 1940*s to 1965. Reported 7-10 percent loss prior
to late 1940's (and beginning of 1080 use) and 2-4 per-
cent loss up until 1965.
Reynolds and Gustad (1971) - Colorado, Montana, Texas
and Wyoming only. Data from USDA Crop and Livestock
Reporting Service or OSDA Statistical Reporting Service
reports of predator 'losses in these states for 1966-69 .
Loss reports ranged from 3.6-7.9 percent.

q.S. Forest Service (USFS) Estimates - USFS records total
livestock losses during the grazing season. Producers
are as Iced to estimate proportion of losses to various
causes. "The Cain Report" noted that predator losses
for Utah (0.4-1.4 percent for the grazing season)
were consistent with the estimates of Nielson and
Curie (1970) and Morris for entire years.
USDA Statistical Reporting Service Estimates. Reports
of losses to all causes. "The Cain Report" noted that
these data place a "ceiling" on the numbers of losses
attributed to predators. "The Cain Report" mentioned
estimates for three individual states: Utah, usually
9-11 percent (range 7.9-14.9 percent); Idaho* usually
7-8	percent (range 6.1-16.1 percent); Wyoming/ usually
8-9	percent (range 5.4-13.8 percent).
Although "The Cain Report" noted general agreement among the
studies, it indicated that the reported loss estimates may ex-
aggerate the actual predation loss rates for the following reasons:
the scavenging of lambs which died due to malnutrition or birth
defects might occasionally be reported as predator kills? the
killing of weakened animals "that would have died anyway" (p. 45)
could be attributed to predation and not other causes; and the
"current heated predator control climate" (p. 45) caused possible
exaggeration by ranchers of losses to predators. "The Cain Report"
also noted the finding of Nielson and Curie (1970) that extremely
high loss levels were suffered by only a small proportion of Utah
producers, while the majority suffered losses of less than 100
lambs and ewes (combined) per 1,000 ewes.
Cain took the view that some loss to predation would be
expected for any prey species. In the absence of predation, the
impact of other factors causing mortality would be expected to
increase (p. 52). While implying a belief that losses to predators
represent a minor proportion of all sheep lost (e.g., p. 47 and

52), "The Cain Report" concluded that the available data on sheep
losses to predation and other causes were too vague to strongly
support any conclusion/ adding:
... The ambiguity only points up the need
for careful research to determine the true
magnitude and nature of sheep losses and the
effectiveness of predator control in reducing
them. And it raises a serious question as to
the benefit derived from control operations.
(p. 52).
B. Sheep Loss Information Since 1972
The most comprehensive review of studies on livestock preda-
tion conducted since 1972 is the 1978 USDI Pish and Wildlife Ser-
vice report entitled "Predator Damage in the West: A Study of
Coyote Management Alternatives." Tables IV-1 and IV-2 are adapted
from this report and include other information and studies not
available in 1978 when the USDI report was completed. Table IV-1
provides a brief description of the characteristics of the loss
studies conducted since 1972. Table IV-2 summarizes the sheep and
goat losses reported in these studies.
1. Direct Field Studies
Direct field studies (also called biological studies) provide
the most reliable method of estimating minimum and maximum livestock
loss to predators. These studies are conducted by biologists
who intensely search for dead animals and verify cause of death.
Minimum losses to predators are the number of animals found and
verified as predator killed. However, even with intense searches,
not all lost sheep can be found. Maximum loss to predators then
is the number of animals found and verified as to cause of death
plus the number of unaccounted animals^ Estimates of predation
rates from studies of this type are usually somewhat lower than are

Table IV-1. Summary of Ewe, Lamb and Goat Loss Studies - 1972 to 1981
No.	Source
Predator Control
1. Slielton (1972) Texas
2.	Klebenow &	Nevada
McAdoo (1976)
3.	Nass (1977)	Idaho
4.	Taylor et al. Utah
5.	Tigner & Larson Wyoming
6. Brawley (1977) Montana
7. Henne (1975, Montana
8. Munoz (1977) Montana
9. McAdoo &
Klebenow (1978)
10.	Delorenzo &
Howard (1976)
11.	Robel et_ al.
Hew Mexico
1967-71 By experiment station
1973	Yes
1973-75	Yea
1972-75	Yes
1973-75	Yea
1976	Control research
1974	Partial
1975	Partial
1976	No
1974-75	No
1975-76	Yes
More protection
than ranchers
could afford
Range lambing,
range grazing
Shed lambing,
range gracing
Mostly shed lamb,
all range grazing
Range lambing,
range grazing
Shed lambing,
fenced pastures
Shed lambing,
fenced pasturea
Shed lambing,
fenced pasturea
Range grazing
Fenced range
Fall lambing
No herders. Research flocks
with good records.
Migratory range with herdera.
Shed lambing operations with
Ilerders and no herders.
Typical range aheep operations.
Range lambing operations with
No herders. High loss ranch
with loss control experiments.
Study without herders and pred-
ator control.
Study without herders or pred-
ator control.
Study with herders but no pred-
ator control.
Study without herders or pred-
ator control.
Mostly night confinement. Rela-
tionships between husbandry
practices and losses were

Table 1V-1. Summary of Ewe, Lamb and Goat Loss Studies -
1972 to 1981 (Continued)
Predator Control
2.	Reynolds &
Gustad (1971)
3.	Early et_ al.
(1974a, bT~
4.	Nesse ejt^ al.
5.	Nesse al.
6.	Nesse et^ al.
7.	deCalesta
8.	Gee et al.
9. Heduna (1977) Kansas
20. Walther et_ al. Texas
California 1974
California 1974
Some ranches
Some ranches
Some ranches
Some ranches
Only by 25% of
1967-78 Yes
21. Terrill (1980) 33 States 1958-79 Yes
All types
Range grazing
Fenced pastures
Penced pastures
and range
All types
Fenced pastures
All types
Fenced pastures
Various types
All types
Herder and no herders.
Economics of predation studied.
Herders. Sample range sheep
No herders. Monitored fiscal
1973	losses.
No herders. Monitored fiscal
1974	losses.
Herder and no herders. State
No herders. Losses charac-
Herders and no herders. Survey
of western sheep industry.
No herders; 80Z of sheep penned
at night. Relationships between
husbandry practices and losses
were studied.
Herding practices not identified.
State Survey.
Compilation of predator loss

Table IV-2. Summary of Ewe, Lamb and Goat Losses to Predators -
1972 to 1981

All Causes

All Predation
Ooyote Predation



Field studies with control


I 9.3}b>






C 9.01
























Field studies with various levels of control or without control.




















Table IV-2. Summary of Ewe, Lamb and Goat Losses to Predators
1972 to 1981 (Continued)
Study		All Causes			All Predatlon	Coyote Predatlon
No.	Type	Ewes	Lambs Ewes	Lambs Ewes	Lamb
Questionnaire Surveys









2.5 4





a^Data adapted from USDIr 1978 for all studies with the exception of No's 11, 20# and 21.
No two studies reported losses in the same way. Additional calculations were necessary to bring the data into a
consistent foratat. These calculations were made by wildlife research biologists of the Section of Predator Damage,
Denver Wildlife Research Center. Dashes are shown in the cells for which estimates are not available.
b|Losses of ewes and lambs combined; data were not reported separately for ewes and lambs.
®>Data include spring and suimer only. Losses in fall and winter were negigible.
"J Data for only 113 days in sunmer.
^'Designated as "stock sheep" rather than ewes.
Losses were reported as percent of total loss and not percent of total ewe and lamb inventory.
NOTE: Study numbers refer to Table IV-1.

predation rates reported in mail or interview surveys. Direct
field studies are expensive and only a small fraction of the sheep
industry has been sampled via this method. Due to the limited
sample, it is not known to what extent predation rates recorded in
direct field studies represent loss rates for the entire western
sheep industry.
When interpreting the results of direct field studies (studies
1 through 11, Tables IV-1 and TV-2), it is necessary to appreciate
the many factors affecting reported loss statistics. Predator
control, for example, varies widely from place to place and can
lead to misconceptions about the actual extent of "predator control"
being practiced. In some of the studies, predator control was
withheld or restricted on the cooperator's ranch, although some
control may have been conducted on neighboring areas (USDI, 1973).
In other studies, normal legal means of predator control were
utilized (control) while in other studies some predator control
methods were used (partial control).
Field studies conducted since 1972 with predator control in
effect showed losses from all causes ranging from 4 to 10 percent
for ewes and 11 to 18 percent for lambs. Losses attributed specif-
ically to coyotes were 0 to 2.3 percent of ewes and 1.2 to 7 per-
cent of lambs (Table IV-1, studies 1 through 5). Field studies
with partial control or where predator control is practiced at
varying degrees (studies 6 and 7) showed losses from all causes
ranging from 6.8 percent of ewes to 26.5 percent of lambs. In
these studies, losses attributed to coyotes ranged from 1.2 per-
cent of ewes to 12.9 percent of lambs.
Several intensive field studies in which predator control was
restricted (studies 8 through 11) were conducted since 1972.
Two of the studies lasted two years, and the third study (study
10) was limited to one summer. In the studies covering the entire
production year, predators killed 0 to 8.4 percent of the ewes
and 12.1 to 29.3 percent of the lambs. These loss rates are gener-
ally higher than those recorded with predator control in effect
(studies 1 through 5). However, other than the presence or absence
of predator control, the small number of studies preclude further

analyses of differences between the. two groups.
In 1981/ Robel et al. (study 11, Table IV-1) reported results
from a 15-month field study conducted from June 1975 through August
1976 in a nine-county area of south central Kansas. They reported
that death losses to canine predators comprised 16.8 and 12.5
percent of stock sheep and lamb losses, respectively. Annual
losses to predators in the nine-county area were 0.9 percent of
the stock sheep inventory and 0.9 percent of lambs. According to
the authors, these losses were about one-fourth the estimated
predator loss for Kansas as reported by USDA-ERS from mail survey
data. Robel et al. used a system involving personal contact with
all cooperating sheep producers coupled with monthly reporting.
Robel believed that this system reduced the reliance on the pro-
ducers' memory and thereby greatly increased the accuracy and
reliability of their predator loss data.
2. Mail and Interview Surveys
Results from mail surveys are based on responses from a sample
of producers to questions on production and marketing practices,
predation problems and predator control. Interview surveys gen-
erally cover the same information but are based on information
collected during personal question and answer sessions"between
ranchers and researchers. Surveys offer the advantages of covering
a much larger sample size and a broader cross section of producers
with predator losses than do single ranch biological studies.
Unless carefully designed, results from mail and interview
surveys appear to have the greatest uncertainties and potential
for bias. Producers generally devote less time to daily searches
for missing sheep than do researchers involved in direct field
studies. Dead livestock found by producers can be identified as
predator killed, but the actual fate of missing animals is often
Mail and interview surveys conducted since 1972 are summarized
in Table IV-1 (studies 12 through 20). In most of these studies,
predation rates were similar to predation rates in direct field

studies where predator control was practiced (i.e., 0.1 to 2.5
percent of ewes and 0.7 to 8.1 percent of lambs).
One study however, reported higher losses than others but was
much more comprehensive because it covered half of the sheep pro-
ducers in the West. In this study, based on farmers and ranchers
surveyed in 1974, Gee et al. (1977) (study 18) reported on "Sheep
and Lamb Losses to Predators and Other Causes in the Western United
States." They reported that rates of loss to coyotes varied consid-
erably among farmers and ranchers. Some had minor or no losses to
predators while others had very high losses. Overall losses
approximated more than 8 percent of all lambs born and more than
2 percent of the inventory of adult sheep. USDI is in general
agreement with these rates of loss estimating annual nationwide
predation loss averages of 4 to 8 percent of lambs and 1 to 2.5
percent of ewes (Andrus, 1980) .
The relatively high rates of predation reported by Gee et al.
(1977) as compared with other studies during this period are not
easily reconciled. In both California and Kansas, for example,
USDA mail questionnaire surveys have generally reported predator
loss rates that are higher than direct field studies that validate
death loss (compare Nesse et al/, 1976 and Robel et al., 1981).
Regarding the reliability of the ERS estimates, Gee et al. (1977)
wrote "... the total number and proportion of producers affected
(without regard to loss level) by coyote predation are probably
quite realistic, since most producers know whether or not coyotes
are preying upon their herds. The numbers and percentages of
sheep and lambs lost to coyotes and the number of producers with
different levels of loss must be considered more cautiously since
the degree of producer judgment is higher."
Terrill, 1980, estimated sheep and lamb losses to predators
in the united States from 1958 through 1978. These estimates were
based on available USDA statistics and other reports. As reported
by Terrill, sheep and lamb losses averaged 4.61 percent of inven-
tory plus lamb crop for the period 1958 through 1971. During this
period, the loss rate increased 32 percent from a low of 4 percent
in 1958 to the high of about 5.28 percent in 1971. During the

next seven years from 1972 through 1978, the loss rate averaged
6.22 percent ranging from 5.65 percent in 1972 to 6.83 percent in
1978 for an increase of 35 percent over the 1958-71 annual average.
A survey of Texas producers also shows an increase in preda-
tion rates. Texas reported the results of a survey of sheep and
goat producers, based on mail questionnaires sent in conjunction
with the January 1, 1979/ USDA Livestock Survey. Reported death
losses of all sheep and lambs to predators comprised 58 percent of
all deaths in 1978 compared to 25 percent in 1967. In 1978, coyotes
were reported as the largest single cause of death comprising 24
percent of sheep and lamb deaths. Other leading causes were eagles
20 percent, bobcats 18 percent and weather 10 percent. In 1978,
sheep losses to predators in Texas reported in the Texas survey,
comprised about 30 percent of all sheep losses compared to 15
percent in 1967. As reported by Texas, predators caused 67 percent
of all lamb losses in 1978 compared with 40 percent in 1967.
Adult goat losses followed similar trends with 54 percent of total
losses to predators in 1978 compared with 30 percent in 1967.
About 79 percent of kid losses in 1978 were attributed to predators
compared to 59 percent in 1967.
Finally, USDA statistical publications, based on producer sur-
veys, report death loss from all causes of all sheep and lambs.
On the national level, death loss to all causes has increased
since compound 1080 was first available in the late 1940*s. During
the next three decades, death losses from all causes have steadily
increased ranging from a low of about 13.1 percent in 1954 to a
high of 16.7 percent in 1975 (Table IV-3).
In 1950, losses of adult sheep comprised about 60 percent of
total sheep and lamb losses. During the next 22 years, the number
of adult sheep losses (as a percent of all losses) steadily de-
clined but remained at more than 50 percent of all losses. In
1972, lamb losses (on a percentage basis) were more than sheep
losses for the first time.* In that year, about 51 percent of all
* The cancellation of compound 1080 in 1972 would not have
seriously affected lamb losses in.1972 since most compound 1080
use in that year would have been completed prior to the
cancellation action.

death losses to all causes were lambs. In 1977/ death losses
reached a peak, comprising about 16.4 percent of stock sheep on
hand January 1. Lamb losses accounted for more than 56 percent of
total losses in 1977 (Table IV-3) .
3• Other Information
In addition to direct field studies and mail and interview
surveys, individual livestock producers, livestock producer
organizations, state and local governments and other interested
parties have provided information pertaining to livestock preda-
te ion and comments regarding the need for compound 1080. This
information was provided in testimony and written comments sub-
mitted to the Agency as the result o£ informal hearings held in
late July. The following summarizes the information provided.
a. Livestock Producer Testimony and Comments
Livestock producers (primarily sheep producers) from 12
states provided information during the Agency's informal hearings.
Producers with diverse production, management and predator con-
trol practices provided data describing livestock losses due to
Because producers reported many categories of livestock
loss information, the basis for comparing livestock losses (e.g.,
lambs lost as a percent of total lamb crop, lambs lost as a
percent of total number of ewes, etc.) is not always consistent
between each producer. In addition, livestock losses were often
reported by producers as total losses to all causes rather than
differentiating losses to specific cause. Producers generally
indicated that coyotes were the primary cause of losses due to
Arizona, Idaho, Oregon, and New Mexico livestock producers
responded with information expressing their support for reregis-
tration of compound 1080 for predator control. These producers

Table IV-3. Death Loss of Sheep and Lambs from all Causes, 1950 to 1973.
Death Loss of all
All Sheep & Lambs Deaths from Sheep and Lambs as
Year	on Hand, Jan. 1	all Causes	Pet, on Hand, Jan. 1.




Source: tJSDA statistical Bulletin No. 522

also indicated that predation problems are increasing despite
improved management practices and the use of legal predator
control methods.
Colorado and Wyoming producers supplied the majority o£ re-
sponses emphasizing sheep losses due to predation. Various Colo-
rado sheep producers reported pre-1972 losses of 0 .4 to 2 percent
of their lambs and post-1972 losses of 3 to 20 percent of their
lamb crop, averaging approximately 11 percent annually. Wyoming
producers indicated that before compound 1080 was used for preda-
tor control, annual lamb losses averaged 10 percent of the lamb
crop. Annual lamb losses of 2 to 5 percent occurred when compound
1080 was available for predator control and increased to 5 to 50
percent (averaging approximately 21 percent) after compound 1080
was cancelled. Montana sheep producers reported annual lamb losses
of 10 to 20 percent since 1972/ averaging approximately 13 percent.
Responses from Kansas, Texas, Nebraska, South Dakota, and
Utah were received from one to three producers in each state.
Producers stated that sheep and lamb losses were generally in-
creasing despite improved management practices and the use of
legal predator control methods. A South Dakota producer indicated
a 14 percent lamb loss in 1979, one Kansas producer lost 15 and 11
of his lambs and ewes, respectively in 1977, while a Texas producer
reported a complete loss of kid goats and a 30 percent loss of
adult goats to coyotes. Two individual livestock producers, Mary
Wintch and John Hotchkiss, provided statistical data on trends in
predation from their operations. The Wintch data included death
losses from all causes with over half attributed to predators.
These data are summarized as follows:
Wintch Livestock Company
Hotchkiss Ranch
Number of Lambs Lost as
Percent of Lambs Docked
Number of Lambs Lost as
Percent of Flock Size
0 .6

Wintch Livestock Company
Hotchkiss Ranch
While limited data are presented prior to 1972/ the reported data
show that losses were generally lower prior to 1972 than in subse-
quent years and/or losses generally have increased during the
In summary, the comments and testimony received by the Agency
from*individual livestock producers indicate that predation losses
(primarily by coyotes) have increased since 1972. They indicate
livestock losses have increased despite intensive management and
uses of available predator control methods. According to livestock
producers, losses to predation combined with adverse market forces
create a situation that many livestock producers consider intoler-
b. Producer Associations
Several livestock producer associations provided oral and
written testimony and/or written comments regarding livestock
losses to predators during informal hearings held in July. A
brief tabular summary of this information follows.
California Wool Growers 1970 to 1980 - 33,553 sheep lost due
Association	to coyote predation.

Idaho Wool Growers
Same as provided by New Mexico Dept.
of Agriculture.
Montana Wool Growers
Some producers suffer significant
lamb losses while others do not.
New Mexico Farm
1980 state-wide survey - lamb losses
due to coyote predation average
4 percent.
Oregon Sheep Growers
Livestock losses have increased
each year since 1080 ban.
Utah Farm Bureau
Lamb losses since 1080 cancellation
range from 8 to 20 percent.
Utah Wool Growers
Association and
Utah-Idaho Farmers
91.9 percent of all predator losses
are due to coyotes.
1978 - 16 percent of all sheep and
lamb losses due to predators.
1979	- 18 percent of all sheep and
lamb losses due to predators.
1980	- 19 percent of all sheep and
lamb losses due to predators.
Wyoming Stock Growers
1974-1980- Total sheep numbers
declined by 29.5 percent/ total number
of sheep lost to predators declined
by 41.2 percent.

Wyoming stock Growers
Associations (con't)
1974 -1.7 percent of total sheep
numbers were lost to predators.
1980 - 1.4 percent of total sheep
numbers were lost to predators.
Wyoming Wool Growers
1965 to 1972 - 4.87-7.93 percent
of all sheep and lambs were lost to
1972 to 1981 - 7.37-10.43 percent
of all sheep and lambs were lost to
predators *
c* State and Lpcal Governments
Like producer associations, several state and local govern-
ment agencies provided oral and written testimony and/or written
comments regarding livestock losses to predators during informal
hearings held in July. A brief tabular summary of this informa-
tion follows.
Montana Dept. of
Historic livestock losses due to
coyotes - 4-8 percent.
Decline in sheep numbers due to
market factors and depredation.
Factors affecting depredation -
1) size of operation, 2) location,
3)	rangeland vs. fans flocks, and
4)	vegetational differences.

Nevada Dept. of
New Mexico Dept. o£
Oregon Dept. of
No loss data provided.
1974 survey - 8 and 2.5 percent of
vestwide lamb crop and adult herd
lost to coyotes, respectively.
1980 - 4 percent of lambs lost to
coyotes in New Mexico.
Summary of several state surveys?
Post-1972 livestock loss rates
approximately double pre-1972
loss rates in Colorado, Montana,
New Mexico, and Wyoming.
Post-1972 livestock loss rates
increased slightly in Idaho, South
Dakota and New Mexico (mail
questionnaires and personal interviews).
Post-1972 loss rates in Texas
indicate a "definite increase"
in predation.
Post-1972 loss rates in Nebraska
"rose slightly" then declined.
1980 livestock losses to predators
Sheep - 4,300*
Lambs - 10,300*
*99 percent of losses due to coyotes

South Dakota Dept. o£
Game, Fish and Parks
No. Increase
Texas Dept. of.
See Texas Survey previously dis-
cussed in "Hail and Interview Sur-
vey Section"
Washington Dept. of
1964-1972 - 3.6-5.5 percent of lambs
lost due to coyotes.
1972-1979 - 7.5-20 percent lambs
lost due to coyotes.
Wyoming, Campbell
Animal Control Board
1972-1981 - Sheep population declined
by approximately 52 percent due to
d. Universities and Other Interested Parties
In addition to livestock producers, livestock producer
associations and state and local government agencies, universities
and other interested parties provided oral and written testimony
and/or written comments regarding livestock losses to predators
during informal hearings held in July. A brief tabular summary
of this information follows.

Oregon State	Animal Nos.	Losses to predation
(compared to baseline period)
(baseline period) 1967-1971 - declined
1972-1975 - (-)13.3%	(-)5.3%
1975-1980 - (-) 6.5%	(-)27.1%
U.S. Dept. of Interior	Livestock losses due to coyotes and
Bureau of Indian Affairs wild dogs have increased annually
since 1972.
International Association
of Fish and Wildlife
Coyote is an important livestock and
wildlife predator that has not been
managed properly since 1080 ban.
Wyoming Financial	Due to predator loss, banks are
Institutions	often reluctant to lend operating
or start-up funds to sheep
4. Summary of Sheep Losses
Many studies of sheep loss to predators have been completed
since 1972. Direct field assessments with normal predator control -
have shown predation of 0 to 2.5 percent of the ewes and 1.3 to
7 percent of the lambs annually. The majority of predation in
different studies was attributed to coyotes.
Questionnaire surveys also show the coyote as the main preda-
tor. Predation losses are unequally distributed both geographical-
ly and among producers, various local or regional surveys report
annual predation loss at 0.1 to 2.8 percent of ewes and 0.9 to 4.7

percent of the lambs. The most comprehensive survey to date (Gee
et al., 1977a) estimated total predation losses of 3.4 and 11.4
percent, respectively, of ewes and lambs in 15 western states in
1974. Some 2.5 percent of the ewes and 8*1 percent of the lambs
were reportedly killed by coyotes. Limited comparisons with other
studies suggest that these estimates are substantially higher than
estimates from other sources (USDI, 1978).
According to USDI (1978)# it is difficult to obtain precise
overall estimates of sheep losses to predators. Each method of
loss assessment has limitations and possible sources of bias. As
reported by USDI (1978)/ losses to sheep producers in western
states during 1972 averaged 4 to 8 percent of the lambs and 1 to
2.5 percent of the ewes. However, average loss rates do not ade-
quately portray the nature of coyote depredation on sheep as losses
are unequally distributed both geographically and among ranchers
and producers (USDI, 1978). USDI concludes: "Truly comprehensive
records of predation losses of sheep*over time are lacking, but
the available estimates point to higher predation rates in the
early 1970's than the 1960's."
Livestock loss data presented by producers, producer associa-
tions, state and local government organizations, and other organi-
zations present a diverse picture of the predation problem. Losses
reported by individual producers are often substantially different
from producer to producer and may vary considerably from loss data
generated from surveys. The nature and diversity of reported
losses tend to substantiate a conclusion that predation is often a
producer-specific problem. In other words, many producers have
no predation problems while others, despite their best efforts to
control predators, suffer very significant losses of livestock.
C. Cattle and Calf Losses
Predators are a problem to cattle producers, but not to the
extent experienced by sheep producers. Pew field or biological
surveys are available to evaluate the nature and extent of the
problem under various husbandry or predator control practices.

USDA-ESCS (1976) conducted a comprehensive personal inter-
view survey on the beef breeding industry in 1975 and as part of
this survey included questions on cattle losses to predators.
USDA-ESCS reported that predators killed less than one-tenth of
one percent of the January 1, 1975 inventory of beef cattle (500
lbs. and over) in the survey populations. Highest losses were in
the Southwest* and most predator losses were attributed to coyotes
and dogs.
USDA-ESCS (1976) reported losses to coyotes in the western
states ranging from 0.4 reported in the Great plains* region to 0.8
percent in the Southwest. The highest loss to all predators, 1.1
percent, also occurs in the Southwest. Calves are particularly
vulnerable to predators in the first 6 to 8 weeks of life but by
weaning time/ the probability of predation is virtually eliminated
(tJSDI, 1978). As with sheep losses, losses of calves to coyotes
and other predators are unequally distributed. In the three west-
ern regions, the proportions of producers reporting any losses to
coyotes ranged from 14 percent in the Great Plains to 26 percent
in the West. Two percent of respondents in the Southwest and one
percent in the West reported losses to coyotes of more than five
percent of calves born (USDI, 1978) .
Losses of calves (less than 500 lbs.) to predators occur at a
higher rate than losses of cattle but are minor compared to losses
to other causes. By comparison, calf losses to theft, disease and
other causes are substantially higher, ranging from 3.6 percent in
the Southwest to 9.1 percent in the Great Plains region (USDA-ERS
1976) .
USDA statistical publications based on producer surveys, report
death loss from all causes to cattle and calves. Based on these
data, on the national level, death loss of cattle and calves from
all causes remained remarkably constant from 1950 to 1971. For
this period, the death loss of all cattle and calves averaged 8.9
See Figure IV-1.

sigura 17-1.
Livestock Production Raglans of che Uaitad S-cases

percent,* ranging from a low of 8 .4 percent in 1966 and 1968 to a
high of 9.5 in 1950 (Table IV-4). Prom 1972 to 1980# the rate of
death loss averaged 11.2 percent, an increase of over 20 percent
from the average for the 1950-72 period.
In addition to the USDA-ESCS survey data and the USDA statis-
tical publication, individual livestock producers, livestock pro-
ducer organizations, state and local governments and other inter-
ested parties have provided information pertaining to livestock
predation and comments regarding the need for compound 1080. This
information was provided in testimony and written comments submitted
to the Agency as the result of informed hearings held in late
July. The following is a brief summary of the information provided
regarding predator loss to cattle and calves.
Texas and Southwestern
Cattle Raisers
1973 survey of 87 percent
respondents indicated problems
with predators.
1977 survey - 1.1 percent of all
calves born are lost to predators;
2.0 percent of survey respondents
reported calf losses of at
least 5.0 percent to coyotes.
Coyotes are known carriers and
potential transmitters of
National Cattleman's
1970 to 1980 - Requests for
Animal Damage Control assistance
in 16 western states
increased 430 percent.
Expressed as percentage of cows on hand January 1.

Oregon Department of
1980 Livestock Losses To
* 99 percent of losses due
to coyotes.
South Dakota Department
of Game, Fish and Parks
118 .2
Oregon State University		Cattle	
Animal Nos.	Loss to predators
(baseline period) 1967-71 - stable
1972-75 - stable	doubled
1976-30 - stable	same as

17-4. Deach Loss
o£ Caccle and. Calves from
all Causes 1950 eo 1980.

Death Loss
from All Causes

as sarceac <
Sand, Jan. 1
Hand Jan.


. 4,232
Inventory of co«s and heifers chae have calved.
Source: Data Resources, lac. 1981.

A. Information on the Efficacy of 1080 and Alternatives
Considered in 1972
In 1972/ the U.S. Environmental Protection Agency followed
the recommendations of "The Cain Report" and a petition by several
conservation groups (Natural Resources Defense Council, et al.,
1972) in banning the use of 1080 and three other toxic chemicals
for the control of predatory mammals. The Administrator described
the benefits of predator control toxicants, in general, as
"speculative and ill-defined" (Ruckelshaus, 1972) . He concluded
There is no reliable data as to the amount of
predator control achieved by the use of these
and that
There is no reliable data as to the loss of sheep
that might occur without a predator control pro-
gram using these poisons... (Ruckelshaus, 1972, p.
In recommending the cancellation of these chemicals, Cain
et al. (1972) concluded that effective resolution of local
predation problems could be achieved by use of methods then
available (shooting, denning, and "the use of more humane traps")
(p. Ill) or by use of procedures that would be developed through
research ("truly specific poisons," repellents, reproductive
inhibitors, live trapping, transplanting, "basic ecological
mechanisms") (p. 111). In addition, the report recommended the
use of extension programs
to encourage livestock producers to solve their

own problems by accepted methods directed to-
ward specific marauding animals, especially
near lambing grounds and closed pastures.
... Extension should not be concerned only with
control; there is a need for assistance and
encouragement in the use of better husbandry
and management practices ... (p. 111).
Many of these conclusions of the Cain Commission were echoed
by persons (cf. Klataska, 1981; Wentz, 1981) testifying at the
information gathering hearings in opposition to reinstatement of
the use of 1080 as a predacide.
The following three sections discuss the use patterns which
have been proposed for the reintroduction of 1080 as a predacide.
Subsequent sections present discussions of alternative methods
for controlling predation and lists of husbandry practices which
have been found to affect livestock losses. As Cain et al.
(1972) had concluded that these alternative control practices
(whether in use then or projected for the future) and husbandry
practices could provide all of the livestock protection needed,
discussions center upon new approaches and recent findings or
interpretations of old data which contradict or define the limits
of this conclusion.
As mentioned in Section III, in reaching the decision to
suspend registrations for chemical practices (Ruckelshaus,
1972), the Administrator used "The Cain Report" (Cain et al.,
1972), "The Leopold Report" (Leopold et al., 1972) and the Nat-
ural Resources Defense Council et al. (1971) petition as major
sources of information. Table V-l summarizes the conclusions
expressed in these documents or in the Administrator's notice
of suspension concerning 1080 use patterns and alternative con-
trol methods. In addition, Table V-l notes whether new data
have become available since 1972 on the various methods and
whether the data may warrant any re-evaluation of the conclusions
reached in 1972. In interpreting Table V-l and for the

Table V-l. 1972 Assessments of Methods for Controlling Livestock Predaticn and
Mditional Information Affecting Current Assessnents of These Methods.
1972 Assessment
New Data?
(post 1972)
Areas for
Ttoxic collar None
Offending animal
Bait station
Effective in local population reduction1,2
"... no reliable data" on benefits3
"Pair" for troubleshooting1
"Very good" as economical methods-
Misused (dosing, density, persons
Wildlife hazard1,3,4
Endangered species hazard3,4
Secondary poisoning1,3,4
Hazard to humans1,3
No new experimental
data cn method
Hobinson (1948)
data re-examined
Single lethal
Less selective than carcass stations
when baits are broadcast1
Sodiun Cyanide
M-44 or
cojote getter
Effective as prophylactic tool1
Poor troubleshooting tool1
"Good" as economical method1
Fairly specific for offending animal2
(M-44 only)
Lack of environmental
Not selective1,2,3
Wildlife hazard3
Human hazard1,3,4
No data on use of
1080 in single
lethal baits to
.control coyotes
New data are on
selectivity of
not-lethal baits
Selectivity and
safety limitations
of ft-44 device

Table V-l. 1972 Assessnents of Methods for Controlling Livestock Predaticn and
Additional Infocraation Affecting Current Assessnents of These Methods.
1972 Assessnent
New Data?
(post 1972?
Other Chemical Methods
(no specific
"Poor" prophylactic tools^
"Very good" troubleshooting tools^
Presumed safety, selectivity for
offending target animals^-
Questionable eccnany, environ-
Areas for
Agents tried have
not proven to be
(no specific
"Very good" for safety, humaneness^
"Good" for prophylactic effects,
lack of environmental impacts^
"Poor" for offending animal
"Fair" for target species
Questionable economy (thought to
be "Good")1
NOne (no new
Mechanical Methods
Aerial shooting
"Very good" troubleshooting tool^-
"Pair" prophylactic tooll
"Good" selectivity for target
"Fair" selectivity for offending
Questionable humaneness^
"Very good" safety to man and live-
stock, lack of environmental impact^
"Poor" as economical methods-
Offending animal
Use greater than
projected in-
Costs have risen

Tfeble V-l. 1972 Assessments of Methods for Controlling Livestock Predation and
Additional Information Affecting Current Assessments of These Methods.
Ground shooting
1972 Assessment
"Good" for troubleshooting/ target
species selectivity, offending
animal specificity1,2
"Very good" for safety to man and
livestock/ lack of environmental
"Poor" as prophylactic method1
"Fair" as economical method1
"Questionable humaneness1
New Data?
(post 1972)
Areas for
"Good" for troubleshooting, target
species selectivity,
offending animal specificity1
"Very good" safety to man and live-
stock, lack of environmental
"Poor" as prophylactic tool1
"Fair" as economical method1
Questionable humaneness1
Effects on losses
to coyotes
Impacts of Andrus'
(1979) policy
Steel traps "Good" for troubleshooting1
Not very selective1,2
"Fair" as prophylactic tool1
Humane when properly used1
"Good" as economical method1
"Fair" in offending animal
Methods of use
Snares	None	Yes	Selectivity
1.	Cain et al., 1972.
2.	Leopold et al., 1964.
3.	Natural Resources Defense Council, Inc., et al., 1971.
4.	RucJelshaus, 1972.

discussions which follow, the reader should note that the term
"prophylactic" refers to attempts to reduce the incidence of
predation by reducing the predator population in a general area.
The terms "troubleshooting" and "corrective" control refer to
attempts to reduce predation by removing the specific individual
predator(s) responsible for the damage.
B. Efficacy of 1080
1. Bait Stations
The most common method of use of compound 1080 prior to its
being banned as a predacide was in the impregnated bait station.
Large portions (50-100 lbs.) of animal carcasses (generally horses
or sheep) were given multiple injections of a sodium fluoroacetate
solution at closely spaced intervals. These carcasses were then
placed on rangeland in efforts to reduce coyote populations.
While this use pattern was implemented without being subject to
intensive quantitative investigations of its efficacy and
safety (cf. Balser, 1974), Robinson(1948, 1953) reviewed the
method, developing a set of use guidelines which he believed
would minimize hazards to non-target species' populations while,
at the same time, providing effective coyote control. These
guidelines have been adopted with few modifications by several
groups which have been concerned with reinstatement of this use
pattern since 1972 (e.g., ASTM, 1976; Wyoming Department of
Agriculture, 1977). This section discusses the elements of the
use guidelines which may affect the effectiveness of this use
According to Robinson's (1948) scheme, 1080-treated car-
casses would be placed at densities of no more than one station
per 36 square miles.* Due to the extensive home ranges of
coyotes, such placement densities were believed to provide effec-
tive target species exposure. The baits were to be deployed
One bait per township section.

during the winter and early spring when conditions would favor
scavenging by coyotes, preservation of the freshness of the
bait and# therefore) the "drawing" power of the carcass. Kill
o£ coyotes at this time of year would, according to Robinson
(1948), provide efficient population reduction by taking coyotes
during their own reproductive season. Coyote reduction at this
time was also thought to benefit the sheep maximally since it
would be achieved prior to the lambing season and the movement
of sheep bands into the higher range areas.
Prior to the 1972 cancellation, bait3 were dosed at a rate
of 1.6 grams of 1080 per 45 kilograms (100 lbs.) of eviscerated
livestock carcass. The ASTM (1976) recommendations for use of
1080 in bait stations direct a similar dose level. These guide-
lines, however, call for placement of stations at an average of
one per township (36 square miles) but state that "in some areas,
more frequent placement may be required" (ASTM, 1976, p. 11).
Proposed state registrations for this use pattern (e.g., Wyoming,
1981) generally follow the ASTM directions (See Table III-l).
Connolly (1981) has speculated that the dosage level for bait
stations could be reduced without diminishing effectiveness.
The perceptions of the EPA. Administrator and the authors of
the reference documents given major attention in' 1972 concern-
ing this use pattern are summarized in Table V-l. The method
was perceived to be effective in reducing local coyote popula-
tions, but Robinson's (1948) data suggesting that these reduc-
tions in predator densities markedly reduced lamb losses to
predators apparently were not accepted.* The prevailing viewpoint
in the documents given primary weight in 1972 (Cain et al.,
1972; Leopold e_t al., 1964; Natural Resources Defense Council
et al., 1971) was that any possible benefits derived from this
* "The Cain Report" did discuss non-target hazards of 1080
and thallium sulfate that were noted by Robinson (1948) , indica-
ting that the panel was aware of the report. "The Cain Report"
did contain an analyis by Wagner, one-of the commission members
(cf. Wagner, 1972; 1975), which purported to show some depressent
effects on coyote numbers during the period of operational use of
1080 in western states. The impacts of these possible effects
(see below) on total livestock losses were not found to be clear-
cut. (See also the testimony of Bourret, 1981.)

use pattern were outweighed by ethical considerations (e.g.,
"humaneness") and hazards to man, endangered species and other
wildlife (see Table v-1 and Section VI). Patterns of misuse
noted for this use pattern included excessive dosing of carcasses,
placement of carcasses at densities exceeding Robinson's guide-
lines, placement by unqualified or unauthorized personnel and
failure to collect baits in the spring (see Table V-1 and Section
VI) .
The validity of Robinson's conclusions concerning the meth-
ods for safe and effective deployment of 1080 bait stations has
not been tested in a detailed, quantitative manner. Leopold et
al. (1964) accepted the logic of placing stations at one per 36
square miles as a means for controlling coyotes while not ser-
iously disturbing non-target populations but did not cite any
supporting data.
By using data on total known takes in the federal program
for animal damage control and on manpower used in these programs,
Wagner (1972) has developed an index which he believes is an
approximate reflector of population levels for various species
of predators. Using this index, Wagner has assessed the effects
of the use of 1080 on predator populations. Because Wagner has
revised his original assessment of the effects of coyote popula-
tions on sheep losses (Wagner, 1975), his original and subsequent
evaluations are discussed at length.
By dividing the total number of coyotes known to be taken in
the federal control program in a given state in a given year by
the number of man-years expended in the program in the same
state for the same year, Wagner derived the statistic "Coyotes
Killed per Man-year of Effort." When coyote numbers were high,
Wagner reasoned, these animals would be more easily taken than
in years when their numbers were low. Wagner felt that, in the
absence of thorough analysis of the control program's effects,
this measure of the program's efficiency could "provisionally be
used to indicate changes in the populations of coyotes and other
carnivorous mammals" (Wagner, 1972, p. 9).

Wagner assembled data £rom the 1930*s until 1970 or 1971,
depending upon the state in question. Although data were missing
£or some years in some states/ a fairly continuous series of
values was assembled. Wagner divided this series into a "pre-
1080" period and a "1080" period. The pre-1080 period ran from
the earliest available record for a state until the year before
1030 (along with the "Coyote Getter"—see discussion of M-44
device) was introduced for operational use in the state.* The
1080 period extended from the year of 1080 introduction onward.
Since Wagner1s data were originally assembled for inclusion in
"The Cain Report#" he reported no data subsequent to the banning
of 1080 as a predacide. USDI (1979a) has reported data on
¦Coyotes Killed per Man-Year of Effort" for the years since
1972/ thus adding a "post-1080" period to this "accidental"
Analysis of the figures on coyotes killed/man-year reported
by Wagner (1972) for eight western states and supplemented by
user (1979a) for seven of these states indicates two general
patterns o£ results:
1.	Coyotes/man?-year sharply declined slightly
before or after the onset of the- 1080 period/
remained depressed during the 1080 period (1948
or 1949 to 1972) and increased during the post-
1080 period (1973 to the present). These trends
were seen in four northerly, mountainous states:
Idaho, Montana/ Utah and Wyoming.
2.	Coyotes/man-year declined after the introduction
of 1080/ but these declines did not persist.
Changes in the index in Arizona/ Colorado, New
Mexico* and Texas appeared to be more closely
associated with cyclical phenomena independent
* in the states for which Wagner compiled data, 1080 was intro-
duced as an operational tool in 1948, 1949 or 1950.

of changes in control methods used. In the
three states (Arizona, New Mexico and Texas) for
which post-1080 data were reported by USDI (1979a)
these cyclical fluctuations appear to have continued
following the banning of 1080.
Wagner (1972) attributed the differences between the more
northerly and more southerly states in the response of the meas-
ure "Coyotes Killed Per Man-Year of Effort" to a perception that
1080 was more effective in the northern states and to information
that, perhaps as a consequence of that perception, more 1030 was
used in the northern states than in the southern ones. According
to this interpretation, the data on coyotes/man-year from these
eight states support the notions that the use of 1080 can sup-
press coyote populations and that the degree of coyote population
suppression achieved is proportional to the amount of 1080 used.
Wagner (1975) stated that he believed that reductions in coyote
populations were correlated with reductions in sheep losses.
Because these data provide much of the available quantita-
tive support for the efficacy of the 1080 bait station as a
prophylactic control agent and, in conjunction with some surveys
of sheep losses, support for the general effectiveness of prophy-
lactic • control Wagner's interpretations merit critical review.
It should first be noted that the changes in coyotes/man-year
noted by Wagner (1972) are corroborated in certain areas of
certain states by trapline surveys (Linhart and Robinson, 1972)
and by numbers of coyotes bountied (Wagner, 1972) .
Several factors could operate to make the index "Coyotes
Killed per Man-Year of Effort" misrepresentative of population
levels or of the efficacy of 1080 bait stations. Such confounding
factors could include the contributions of other methods to the
total number of coyotes killed and differing contributions of
various methods to the coyotes/ man-year index, either through
different killing efficiencies or different efficiencies of
having kills detected.

The data on coyotes/man-year are composites oi the efficien-
cies of all methods used. Methods which have known takes of
relatively large numbers of coyotes or which involve expenditures
of relatively large amounts of employees' time exert major in-
fluences on this composite measure. The methods used over the
periods covered by the Wagner (1972) and the USDI (1979a) data
include traps, snares, ft-44's, "Coyote Getters," gunning (from
ground or air), denning, and toxic baits laced with strychnine,
thallium sulfate or 1080.* The victims taken are readily located
and, therefore, readily tallied for traps, snares, shooting and
denning. For quick-acting toxicants such as sodium cyanide (in
the "getter" or the M-44) and strychnine, a larger proportion of
victims would be expected to be located than for slow-acting
poisons such as thallium and 1080.** Over their periods of use,
the contributions to the total known kill of coyotes and to the
total actual kill of coyotes would be very similar for some
methods and very different for others.
The different methods also vary in the number of known kills
(or takes) per man—year of use (USDI, 1978). For example, in
fiscal year 1976, the composite efficiency of the U.S. Fish and
Wildlife Services' animal damage control program was 204.5 coyo-
tes taken per man-year of effort. Co.yotes/man-year values for-
individual techniques ranged from highs of 989.4 and 426.2 for
gunning from helicopters and fixed wing aircraft, respectively,
to lows of 70.3 for snares and 113.5 for scdium.cyanide in the
M-44 (USDI, 1978). For fiscal year 1976, denning, hunting with
dogs, ground shooting and aerial shooting produced coyotes/man
year values above the composite efficiency while traps, M-44's
and snares had efficiencies below this figure. There are no
* Although the mix of methods used by OSDI personnel varied
over the years, Wagner (1972) assumed that, with the exception
of 1080, the various methods "each compensated for the other in
its respective period of use" (p. 10).
** The difficulty of locating coyotes believed to have been
poisoned by 1080 has been mentioned and discussed in many sources
(e.g., Robinson, 1948; Hegdal et al., 1978; Connolly, 1980).

known data giving coyotes/man-year values for use of the 1080
bait station. I£ such data were available, Wagner's (1972)
conclusions could be assessed more accurately.
Since animals killed by 1080 frequently are not located, it
is possible that a low known coyotes/man-year value was obtained
for use of 1080. Since cancellation of 1080 was followed by an
increase in the use of aerial gunning, the apparent increase in
coyote populations as measured by "Coyotes Killed per Man-Year
of Effort" may merely reflect the influence that changing control
methods has on this index. The same effect may have operated
between the "pre-1080" and "1080* periods as well. It is note-
worthy that the greatest and most enduring depressions in the
index following the introduction of 1080 as an operational tool
occurred in the states which used the most 1080. This finding
is consistent with the interpretation expressed in this paragraph
as well as with that of Wagner (1972).
Although Wagner's (1972; 1975) interpretations of his data,
as supplemented by USDI (1979), may be correct, the data of
Robinson (1948) provide the clearest support for the effectiveness
of 1080 bait stations in reducing predation loss. A1 though
Robinson's data were limited in validity by his use of year-to-
year comparisons of loss figures obtained through interviews with
producers, the local reductions in losses to predators reported
were of such magnitudes (3S, 98 and 99 percent) that it is diffi-
cult to conclude that subtle biases produced all of the apparent
effects. As noted above, these data were known to the Cain
Commission but were not discussed in the report's section on
coyote control and sheep losses.
Lynch and Nass (1981) have recently summarized loss data on
national forest service lands from the years 1960 to 1978.
Percent loss was significantly inversely correlated with numbers
of 1080 stations used (1960-1972), but percent loss was also sig-
nificantly inversely correlated with the numbers of sheep grazed.
The degree of relief from predation which would result from
reinstitution of this use pattern is difficult to predict.
Despite Robinson's (1948) data, the benefits of the old registra-

tion have been hotly contested. Allegations of misuse and exces-
sive secondary coyote kill3 by 1080 used in rodenticide programs
(e.g., Natural Resources Defense Council et al., 1972) suggest
that more coyotes may have been taken by 1080 prior to 1972 than
would be taken under a reinstituted and tightly controlled regis-
tration of this use pattern. The use of 1080 bait stations in
Wyoming in 197S-1977 (cf. testimonies of Crosby/ 1981; Bourret,
1981; and Strom, 1981) was apparently not accompanied by careful
monitoring of effectiveness of the baits in controlling coyotes
or reducing livestock losses.
2. Toxic Collar
The toxic collar was developed by Mr. Boy McBride, of Alpine,
Texas, in the early 1970's as a way of delivering toxic material
to coyotes which prey upon sheep (Connolly, 1980). The first
collars contained sodium fluoroacetate (1080) in vessels posi-
tioned in the throat region, a common locus of attack for coyotes
preying upon sheep (Timm and Connolly, 1977). A coyote attacking
collared sheep would, according to toxic collar theory, inadver-
tently rupture the collar, causing toxicant to enter the mouth.
Since coyotes are extremely sensitive to 1080 (LD50 about 0.12
mg/kg body weight—Connolly, 1980) , even a small amount of suffi-
ciently concentrated 1080 solution can be fatal. With the
cancellation of 1080 as a predacide in 1972, domestic devel-
opment of the 1080 toxic collar stopped. The use of 1080 in
the toxic collar was not mentioned in the decision to suspend
registration of 1080 as a predacide.
Experiments with toxic collars were begun in 1974 by the
Fish and Wildlife Service, USDI (Connolly, 1980). Because it
delivers toxicant only to animals "guilty" of attacking live-
stock, the collar was perceived as a mechanism for "discrimina-
ting against animals actually doing damage," the type of predator
control recommended by the Cain commission (Cain et al., 1972,
p. 6).* Pish and Wildlife Service field researchers used sodium
cyanide (NaCN) in 1975, diphacinone (DPN) in 1976, and have used
¦ The Cain Commission did not favor the use of toxicants for
such "corrective" control, however.

1080 from 1973 to the present. In 1979, field trials were ex-
panded to include goats.
Following the directive of the Secretary of the Interior to
halt his department's "research or development of potential uses
of Compound 1080" (Andrus, 1979)/ other parties sought experimen-
tal use pecmits for the testing of the 1080 toxic collar. Two
permits have been granted: one to the Texas Agricultural Experi-
ment Station (TAES)/ Texas A&M University; and the other to the
New Mexico Department of Agriculture (NMDA). The USD! permit
for the 1080 collar has been renewed, but the only field studies
being conducted under it are several trials on mohair goats in
Texas in cooperation with the TAES and one trial in Idaho where
collars have been used on sheep since 1978. The toxic collars
being used are manufactured by Mr. McSride, who also carried out
field trials in Texas in 1978 under contract with USDI.
Toxicants Used in Collars. The USDI's experiments with
neck collars for sheep began in 1974 with NaCJ. Over the first
three years of research/ the compounds screened with captive
coyotes in one-hectare enclosures included NaCN, DPN, mandeloni-
trile, 4-aminopyridine, phosphamidon, and 1080 (Connolly et al.,
1978). Three of these compounds/ NaCN, OPN and 1080/ have been
tested for effectiveness in the field (Connolly et al'.,' 1973;
USDI/ 1979b; Connolly/ 1980). At present, methomyl is being
screened for use in toxic collars (Connolly/ 1981).
Collar Designs. During the course of experimental study of
this method of coyote control, several different collar designs
have been developed. An early design tested by CSDI was a bulky
polyvinylchloride (PVC) collar which included 10 packets each
capable of holding 50 ml of liquid (Connolly et al., 1978).
This design was used in the initial NaCN field trials. Modifi-
cations of this design included reducing the numbers of packets,
altering the thickness of the packets, substituting other mater-
ials (e.g., different plastics, rubber) for PVC, altering fill
valve design, and shifting from nylon to velcro straps for attach-
ing collars to animals. Collar color was changed from white to
black. The major reasons for design changes were the needs to:

1} contain the toxicant adequately prior to attack; 2) minimize
chances that collars would be lost or displaced prior to attack;
and 3) maximize chances that attacking coyotes would be poisoned
(i.e., collar designs should neither repel coyotes nor cause
them "inadvertently" to miss or £ail to puncture the collar or
to receive a sublethal dose) . Some design modifications have
also served to promote safe use of the collar both for the person
handling it and the sheep or goat wearing it (Connolly et al.,
1978; CJSDI, 1979b).
The design currently being tested with 1080 consists of a
one-piece black rubber body which is fastened around the neck of
a sheep or goat by means of two Velcro or elastic straps. The
body is divided into two reservoirs. The reservoirs are loaded
with toxicant by hypodermic injection. Two si2es of collars of
this design are now being used: a smaller size with reservoir
capacities of about 15 ml (30 ml per collar), and a larger size
with reservoir capacities of 25-30 ml (50-60 ml per collar,
Connolly, 1980) . The larger size is used on larger lambs and
kids and, occasionally, on adult sheep or goats.
Methods of Collar CTse. Because of expense, labor require-
ments and, in some cases, safety considerations (NaCN), research-
ers have not equipped all animals in a given livestock operation
with toxic collars. Instead, researchers have attempted to
direct coyote predation toward small numbers of animals which
have been collared and placed away from the much larger numbers
which have not. In order to influence the coyote's selection of
victims, biologists have placed collared animals in regions
where predation has occurred in the recent past and have moved
most of the uncollared livestock to other locations which were
felt to be safer from coyotes.
In USDI's NaCN trials, researchers attempted to direct preda-
tion toward individual collared lambs tethered at points along
"routes habitually traveled by sheep-killing coyotes" (Connolly
et al., 1978, p. 198). This approach had led to collar punctures
and coyote kills in pen tests but proved to be unsuccessful in

the field. The approach failed not only because relatively few
attacks were directed at the tethered lambs, but also because
the coyotes that did puncture NaCN collars apparently were not
Despite the apparent failure to kill coyotes, predation
rates "dropped substantially" during the NaCN field trails. The
researchers (Connolly et al., 1978) attributed these declines to
disturbance caused by increased human activity in the test areas.
Following the NaCN trials, flocks were manipulated in dif-
ferent ways in attempts to direct predation toward collared lambs.
Small flocks (up to 10 animals) of collared lambs were placed in
pastures either prior to the arrival of the main flocks or after
the main flocks had been relocated in areas that were presumed
to be safe. These procedures were used for most of the DPN
trials (Connolly et al., 1978). Collars were punctured in the
~PN field trials, and in some cases, declines in predation were
noted after collars were broken. The likelihood that collared
animals would be attacked was strongly affected by the degree of
separation of the main flock from the target flock.
(7SDX abandoned the use of DPN in the toxic collar because
its slow, anticoagulant killing action permitted lethally dosed
coyotes to kill sheep for several days after they had attacked"
collared sheep. Some apparently healthy coyotes shot from air-
craft in the vicinity of the DPN field trails proved to have DPN
residues in their tissues (Connolly, 1979). Some of these aerial
gunning victims would probably have died from DPN poisoning had
they not been shot first.**
* With a fast-acting toxicant such as NaCN, the coyote
carcass would be expected to be found near the site of attack.
No coyote carcasses were found in the NaCN field trails even
though eight collars were bitten.
** These data also show that at least some of the coyotes taken
by aerial gunning are "offending" animals in livestock predation.
That these coyotes could have been dosed through scavenging on
sheep carcasses was unlikely because of the rather rapid post-
mortem clean-up operations used by the researchers in the DPN
field trials (USDI, 1979b; Connolly, et al., 1978).

In the 1080 trials (1978 to the present)/ the sizes o£ target
sheep flocks have been expanded to include more lambs and_ewes
than in the earlier trials. In the 1978 trials (USDI, 1979b),
numbers of collared lambs in target Clocks ranged from 7 to 38,
while the total number of animals in these flocks ranged from 7
to 70. The first goat trial, run in Texas in 1978, had a target
herd of 80 animals, 20 of which wore collars. Target flocks
used in trials since 1978 have tended to be'of 40 or more animals,
with 40 percent or more of the animals being collared young.
Adult animals are occasionally collared, especially on trials
which use goats as the livestock species to be protected.
USDI's field trials with 1080 collars have produced three
major types of results. In one type o£ outcome, predation ceased
after one or two collars were punctured. While it is tempting
to conclude that such triads are clear demonstrations of effec-
tive control, there have been other trials in which predation
stopped in the absence of punctured collars. This second type
of result, which was also encountered in the diphacinone trials
(Connolly et al., 1978), has been attributed to the removal of
the problem coyotes by other means (Connolly, 1980; USDI, 1979b).*
In the third type of finding, killings continue even after a few
collars are broken. In these trials, predation rate often has
been reduced after additional collars were broken and coyotes
were taken by other means as well. In these situations, the
collar was used as a tool in a predator control arsenal. Even in
these cases, however, the action of the collar was corrective be-
cause all coyotes taken with it had demonstrated an interest in
attacking livestock. The collar has not accomplished the total
corrective job in situations in which there were some predators
that did not attack the throat, whether these other predators were
coyotes, dogs, or some other species. (Connolly, 1980; 1979b).
In field trials, it generally has not proven to be
possible to eliminate all methods of control other than the
one under study. The fact that coyotes range over vast areas
and are thus subject to interaction with the interests of
many different humans is a major reason why the elimination
of other types of control is so difficult to achieve.

The collar's selectivity for offending coyotes is its major
advantage. This method is labor intensive to use and can be
very costly when material and labor costs are included (especial-
ly when one considers that the costs of the 'sacrificial" lambs
or kids must be borne by the rancher involved with the use of
the collar). It is believed by some (Texas Sheep and Goat
Raisers Association/ Appendix G, Connolly, 1980) that ranchers
must be permitted to apply collars if they are to be used effi-
ciently. According to this argument, responses to loss situations
would be too slow if ranchers were forced to delay collar applica-
tion until times suited to the schedules of local trappers or
other professional applicators.
The collar is expected to be most useful in farm flack and
fenced pasture situations which facilitate the manipulation and
segregation of target and non-target flocks (USDI, 1979b; Connolly,
1980). It is unlikely that the collar can be used successfully in
typical range operations.
If the collar is introduced for operational use, an accounting
and recovery system may be useful. A system under which ranchers
rent (with deposit) collars from government agencies may provide
sufficient incentives to insure that excess collars are returned
and that searches for lost collars are conducted.
3. Single Lethal Baits
The extent to which this method was used prior to 1972 for
delivering 1080 to coyotes and other species is not known. A
somewhat similar scheme was used to deliver strychnine, but sin-
gle lethal or "drop" baits laced with 1080 received very little
mention in the primary resource documents used by the E?A in
1972 (See Table V-l).
The concept of the single lethal bait is to place coyote-
attractive materials that are laced with enough 1080 to kill one
coyote. In theory, the coyote would consume the entire bait and
would be fatally poisoned (with virtual certainty). The amount
of 1080 in the bait would be carefully controlled so that the

hazards to species less sensitive to 1080 (or larger) than the
coyote would be reduced. The preparer of such baits, then,
would be expected to be very careful to avoid overdosing or un-
derdosing baits. Sublethally dosed coyotes would be expected to
become bait shy (see section [C.3] on aversive conditioning
Bait materials proposed for use are animal products such as
"ground or rendered lard, tallow or other fats" (ASTM, 1976,
p.7). These materials are expected to remain stable in cool
weather but would melt at warmer temperatures. Warm weather is
expected to destroy the baits, leading to the removal and degra-
dation of the 1080 by plants and bacteria, respectively (ASTM,
Two methods of preparation of single lethal baits have been
described (ASTM, 1976). In one method, 1080 concentrate would
be mixed directly with melted bait material. The mixture would
then be subdivided into 10-15 gram portions. In the second
method, pellets or capsules containing a single lethal dose for
the intended target species would be inserted into "preformed
solid baits" (ASTM, 1976, p.9). The amounts o£ 1080 to be used
in preparing single lethal dose baits have been listed as 5 mg
for coyotes, 3 .mg for red.foxes, and 2 mg for gray foxes (ASTM,
1976). The levels listed by ASTM for coyotes and red foxes have
been proposed by Wyoming (1981) and Colorado (1981) in their
applications for'Section 24(c) "Special Local Need" registrations.
In its registration application, Montana (1981) proposes to use
3.6 mg doses for coyotes. (See Table III-l).
In use, single lethal baits are to be placed near "estab-
lished draw stations" (intentionally placed animal carcasses)
or near "preferred travel routes in suitable locations for the
target animals to find them" (ASTM, 1976, p.13). The number of
baits to be placed in one location is left to the discretion of
the applicator, who is expected to weigh various local factors
such as perceived densities of target and non-target species in
determining the proper number of baits to be placed. Bait den-
sities proposed in applications for registration of this use

are summarized in Table III-l.
Varioua procedures have been recociaended for discouraging
bait take by non-target species and encouraging the take by
target species. Seme of these procedures include covering baits
with stones, cowchips, or other objects (ASTM, 1976), burying
baits or elevating them (Linhart et al., 1968; Tigner et al.,
1981). Pish meal or other attractants have been added to baits
to attract coyotes (Tigner et al., 1981). Sonic emitters have
been tested as alternative "draw stations" to livestock carcasses
(Tigner et al., 1981).
The usefulness of these measures in reducing take by non-
target species is discussed in Section VI. Of interest here is
the observation that a bait taken by a non-target organism is a
bait not available for consumption by the target species. Exper-
iments with drop baits that did not contain 1080 (Linhart et
al.* 1968; Tigner et al./ 1981) have indicated that many baits
may be taken by non-target species and that relatively few of
the coyotes collected from baited areas actually consume these
While the information gathering hearings elicited much
testimony regarding the efficacy and safety of single lethal dose
drop baits, very little was offered in the way of evidence.
Glosser's (1981) information concerning the baiting of feral
dogs in Guam is not relevant to the baiting of coyotes in the
western U.S. Data supporting the ASTM (1976) procedures for
using single lethal dose baits were not presented. The Agency
possesses no data on the efficacy of 1080 single lethal dose
baits from either laboratory or field testing. This lack of
data restricts the Agency's ability to assess the possible bene-
fits of this delivery mechanism.
Information currently available does not indicate whether
effective and selective delivery is possible through use of drop
baits. The value of any future field research conducted in this
area would be enhanced by data on densities of species of concern
in the test areas. In the absence of such data, a result showing
that coyotes accounted for nearly all of the baits taken in an

area could mean either that a selective baiting procedure had
been developed or that the non-target species which might be
attracted to the bait placements used were simply not present in
the study area.
C. Other Chemical Methods
1. Sodium Cyanide
Prior to the cancellation decision of 1972, the primary
mechanism for the delivery of sodium cyanide to coyotes was the
"Humane Coyote Getter." This device was embedded vertically in
the ground. When a coyote or other animal tugged on a meat lure
attached to the exposed end o£ the "getter," a gunpowder explosion
was triggered, forcing a cyanide capsule into the mouth and
producing a rapid death. In the M-44 device, a spring ejector
is substituted for the gunpowder mechanism. Although the M-44
device was developed prior to the cancellation of sodium cyanide
as a predacide, the negative perceptions o£ the use pattern in
1972 (Table V-l) were based primarily on the use of the "getter."
Following cancellation of this use, various state and fed-
eral agencies petitioned the SPA for experimental use of sodium
cyanide in the M-44. The use pattern was ultimately registered
(Table V-l)* The "significant new evidence" supporting the regis-
tration of sodium cyanide in the M-44 included the documentation
of a greater degree of selectivity for target species than had
been attributed to the use of the "getter" and the M-44 in 1972.
Nearly three-fourths to all of the animals known to have been
taken with the M-44 in various campaigns have been coyotes, with
canid species (coyote, fox, feral dog) comprising 89 percent
or more of the total known take (Beasom, 1974; USDI, 1978, 1979a).
Reports of limitations and failures of the M-44 device have
been mentioned in recent publications (USDI, 1973) and at the
information gathering hearings (e.g., Levinston, 1981; Barron,
1981; Wade, 1981). Major limitations cited are the 26 use

conditions attached to the registration (USDI, 1973; Rost, 1981;
Wade, 1981) and the inapplicability of the method in winter in
the northern states (Hibbard, 1981; Madsen, 1981; Uhalde, 1981).
Caking of cyanide in the capsules has been cited as a reason why
some animals discharging the H-44 devices are not killed. Me-
chanical failures noted include jamming of devices by dirt and
Despite these problems/ some individuals are now using the
H-44 effectively. The CJSDI animal damage control program is now
engaged in efforts to improve the reliability of this tool.
2.	Repellents and Reproductive Inhibitors
Cain et al. (1972) placed great faith in the notion that
effective chemical repellents and/or reproductive inhibitors
could be developed which would provide effective solutions to
many predator problems. The effects attributed to these methods
were apparently theoretical because specific compounds were not
mentioned. At the information gathering hearings, Havens (1981)
recommended these approaches. Hodder (1981) discussed a "repel-
lent" product applied to sheep but did not disclose its composi-
tion. Ohalde (1981) reported an inability to find an effective
Research reported on both methods before and after 1972 has
yielded generally disappointing results. For example, trials
using diethyl stilbestrol (DES) in drop baits as a sterilant for
coyotes did not produce a technique that could be used operation-
ally to suppress reproduction in this species (Linhart et al.,
1968). Experiments with repellents have failed to identify
chemicals which are consistently repellent to coyotes and do not
harm sheep (Lehner et al., 1976).
3.	Aversive Conditioning Agents
The area of non-lethal chemical control of predation which
has received the greatest amount of research attention since

1972 is the development of conditioned aversions to prey. Pres-
ently* there are no chemicals registered for this use. At the
information gathering hearings, however, representatives of
several conservation and wildlife groups expressed support for
the use of lithium chloride as an aversive agent (e.g., Armen-
trout, 1981; Atkins, 1981; Oungan, 1981; Scott, 1981; Stevens,
1981). Because of this interest and the extensive amount of
recent experimentation on this approach, aversive conditioning
of coyotes is discussed at length.
When an animal becomes ill following the ingestion of any
substance, the animal may subsequently be reluctant to eat that
substance again. Particularly strong conditioned food aversions
result when the ingested substance is new to the animal, when
the substance has .a distinctive flavor, and when the (apparent)
physical discomfort following ingestion is severe (Garcia et
al., 1974). It is not necessary to the development of condi-
tioned food aversions that the ingested material actually be re-
sponsible for the internal malaise, as long as the discomfort
follows ingestion in time.*
* The circumstances under which the formation of a condi-
tioned food aversion has survival value, however, arise when
the ingested material is also the source of the illness inducing
factors. By correctly mentally linking the effect (illness)
with the flavor (for,mammals,- taste stimuli seem to be more
important in this regard than odors), the animal is able to
avoid future poisonings by the same ingested substance or mixture
of substances. While the animals use the conditioned food aver-
sions to their advantage in dealing with toxic materials, the
prime mover behind the elaboration of the process has probably
been the development of toxic "defensive" chemicals by plant
communities. Such toxic plant secondary compounds include some
widely known and used pesticidal agents (e.g., pyrethrins, rote-
none, red squill, sodium fluoroacetate, strychnine, etc.) and
drugs (e.g., caffeine, atropine, opium, quinine, etc.)*
That mammalian pests can form conditioned aversions to toxi-
cants placed by man in control efforts is common experience and
has given rise to the term "bait shyness." This built-in defense
mechanism is of little value to the animal, however, if it has
consumed a lethal dose before symptoms are detected. For this
reason, animals tend to be cautious in sampling new foods.
Nevertheless, many animals are fatally poisoned in pest control
programs or by naturally occurring toxicants. Consumption of
toxic plants can be a significant mortality factor in sheep and
goat raising in the U.S. (Gee et al., 1977; Walther et al.,
1979). Livestock eating Australian or African plants containing
1080 are often fatally poisoned (Aplin, undated; Pattison, 1959).

In recent years, there have been several attempts to exploit
the conditioned aversion phenomenon for man's benefit in animal
damage management (cf., Rogers, 1978). Such endeavors are com-
plicated by the additional variables encountered when moving from
the laboratory to the field and by the fact that it is frequently
necessary to break animals of established feeding habits (as
opposed to conditioning them not to eat a new food) in pest
control applications. Nevertheless, there have been some reports
of success in influencing the depredatory activities of vertebrate
animals through use of the conditioned aversion phenomenon. The
trials involving coyote's selection of prey are reviewed below.
The appeal of exploiting the conditioned food aversion
process to resolve the coyote-sheep problem is that a success-
ful program would spare both prey and predator (Garcia et al.,
1974). Initial studies by Gustavson et al.(1974, 1976) demon-
strated the conditioned aversion phenomenon in coyotes. Four
coyotes fed fresh hamburger laced with lithium chloride, a mild
toxin, became ill (vomited). Four days later, these animals
refused to eat untainted fresh hamburger (Gustavson et al.,
1974). One of three coyotes fed a bait composed of lithium
treated lamb meat wrapped in a woolly hide failed to attack a
live lamb four days later even though all three had killed lambs
two days prior to the exposure to the lithium bait. The two
coyotes that had continued to kill lambs were given a second
experience with a lithium-laced lamb bait followed by an intra-
peritoneal injection of LiCl. These animals refused to attack
live lambs presented four days after the second LiCl treatment.
Results obtained with three coyotes trained to avoid rabbit meat
were essentially similar to the lamb trials: two animals required
two treatments before refusing to attack live rabbits. The
third refused to attack after one treatment. The aversive con-
ditioning did not appear to be permanent for the coyotes con-
ditioned with LiCl after eating rabbit flesh (Gustavson et al.,
1974) .
In a second series of experiments, Gustavson et al. (1976)
reported the conditioning of six coyotes to avoid eating

rabbits after experience with rabbit carcasses or "rabbit-bait"
packages laced with LiCl. Five of these animals required two
experiences with tainted meat. One of these five continued to
kill rabbits introduced into her cage but did not eat them.
Gustavson et al. (1976) also reported results of field trials
in which they attempted to reduce predation by using lithium-
laced baits. The first baits were composed of dog food mixed
with LiCl and wrapped in sheep hides. Subsequently, sheep
carcasses were injected or sprayed with LiCl and moved to bait
station areas or allowed to remain in the spot where the sheep
had been found dead. Feeding on these baits appeared to stop in
March (dog food) and April (sheep carcasses) after about two and
three months of exposure, respectively. Losses for the study
season were compared with the cooperating rancher's loss records
for the three previous years (1972-1974). The authors noted
considerable disagreement between themselves and the rancher
in sheep losses attributed to coyote predation. Using the ranch-
er's records, Gustavson et al. (1976) calculated a predation
rate of 30 percent lower than the average for the previous three
years. Using their own records, the authors estimated a 60 per-
cent reduction in losses to coyotes.
Other field trials in which successful application of con-
ditioned aversions to reduce sheep losses to coyotes have been
reported by Stream (1976a), Ellins et al. (1977), and Gustavson
et al. (1977). Procedures used in these studies were generally
similar to those employed by Gustavson et al. (197S) in that
LiCl was the agent used to induce illness and that the LiCl was
presented in baits consisting of sheep meat wrapped in sheep
hide, injected carcasses of sheep which died on the range, and
dog food wrapped in sheep hide. These studies were undertaken
in Washington (state), California, and Saskatchewan. The Sas-
katchewan data came from rancher use studies involving 19 herds
comprising nearly 22,000 sheep and lambs, and a total combined
land area of nearly 140 square miles (Gustavson et al., 1976) .
Overall, losses to coyotes were reported to be 66 percent lower
in the year that LiCl was used (1976) than they had been in the

previous years. Losses were reported to have been reduced for
14 herds, with loss reductions exceeding 80 percent £or nine of
these. Losses were reported to have increased £or two herds (6
percent and 40 percent), while one herd reportedly suffered no
losses in either year and two store herds lacked loss estimate
data for 1975.
Despite these reported successes of conditioned food aversion
approaches to resolving sheep predation problems, the usefulness
of this technique and the validity of the data reported above
have been questioned. Bekoff (1975) criticized the conclusions
of the authors (Gustavson et al., 1974) of the original demon-
stration of conditioned aversion in captive coyotes on the
grounds that transfer of training from bait to live prey was
established only with difficulty (two of three subjects required
ingestion of two lithium laced baits plus a LiCl injection).
Sterner and Shumaker (1978) noted a lack of appropriate controls
and detailed reporting of procedures in studies reporting
successful application of LiCl to reduce livestock predation by
coyotes. Conover et al. (1977), Lehner and Horn (1977), Burns
(1977), Burns and Connolly (1980), and Griffiths (1978) have all
reported problems with transferring conditioned food aversions
from treated bait3 to live prey. In three of these studies
(Conover et al., 1977? Burns, 1977; Griffiths, 1978), researhers
encountered difficulty in attaining even distributions of LiCl
in injected carcasses and observed that coyotes learned to avoid
treated spots in feeding on such baits.* Conover et al. (1977)
provided their coyotes with intense preconditioning experience
with live and untainted dead prey, procedures which would be
expected to enhance the selectivity of any aversions established
but which also represent a closer approximation of the situation
encountered in the field by the reseacher attempting to break
predators of established feeding habits. The studies with
* Similar problems of toxicant distribution have been men-
tioned for the injection of carcasses wit^i 1080 solution (e.g.,
Natural Resources Defense Council et al., 1971)

captive animals agree that it is possible to establish some
sort of conditioned food aversion in coyotes, but the permanence
of such aversions is debated (Lehner and Horn, 1977), as is
their applicability in field situations (Burns and Connolly,
Griffiths et al. (1978) reviewed prior research with LiCl as
an aversive agent for regulating predation by coyotes. Their
paper includes a detailed critique of the "positive" field re-
sults reported by Gustavson et al. (1976), Stream (1976a), and
Ellins et al. (1977). Their review illustrates the need for
caution in interpreting the results of all predator loss studies.
Griffiths et al. (1978) noted some factors in the
data from the field trials in Washington (Honn Ranch) which
Gustavson et al. (1976) and Stream (1976a) did not take into
account in their discussions of results, although Stream (1976b)
later reassessed his data. Griffiths et al. (1978) noted that
although it was true that fewer lambs were killed and that the
rates of lambs killed per week were lower in 1975 and 1976 than
in 1972, 1973 and 1974, it was also true that the total numbers
of lambs grazed were much lower in 1975 and 1976 than in the
years prior to 1974. The percents of lambs lost in 1975 and
1976 were higher than the values obtained in any of the years
from 1970 through 1974. Griffiths et al. (1978) also noted a
significant negative correlation between the rates of lambs and
ewes killed per week and the numbers of coyotes removed by the
local trapper for the years 1970-76. Changes in husbandry prac-
tices, addition of fencing, and the use of other control methods
concurrently with the LiCl trials were noted as other factors
confounding the Honn Ranch data.
In the Ellins et al. (1977) study (Antelope Valley, Califor-
nia), Griffiths et al. (1978) noted the following weaknesses:
absence of comparable loss data from other regions or other years
(in which LiCl was not used); absence of information concerning
concurrent use of other coyote control methods in study area; and
inadequate explanation for the continued take of LiCl baits after
killing of lambs had ceased. Griffiths et al. (1978) also

speculated that other control methods (e.g., traps, snares,
shooting, 1080 injections in fresh kills) may have been used
along with LiCl in some o£ the study areas in Saskatchewan
(studied by Gustavson et al., 1976). The reviewers acknowledge
a possibility that LiCl used alone may have produced loss reduc-
tions in some of the Saskatchewan study sites, pointing toward a
need for more tightly designed and controlled studies of LiCl in
the field. The authors of a field study conducted recently in
Alberta (Bourne and Dorance, 1981) have concluded that LiCl
baits did not reduce coyote predation on sheep. In this study,
the effects of LiCltreated and placebo baits were compared.
Griffiths et al. (1978) indicated that the development of an
effective bait (i.e., one not subject to uneven toxicant distri-
bution) and the determination of the rate of extinction of an
acquired aversion are essential for the assessment of the con-
ditioned aversion approach as a predation control tool. These
authors also noted that field trials of LiCl induced aversions,
like those of any other method:
. . . must not only demonstrate that predation would have
occurred in the absence of the treatment, but also, that
any observed reduction in predation resulted from the
experimental treatment and not from other causes.
(Griffiths et ad., 1973, p.193)
These performance criteria also have not been met in studies
of other predator control methods. The criteria usually "settled
for" with lethal and/or capture methods are the demonstration
that the method will kill (or capture) coyotes and the compila-
tions of data which indicate the degree of selectivity of the
method for target organisms (variously defined as "offending"
animals, coyotes, canids, carnivores, etc.—cf., Cain et al.,
1972; USD!, 1978, 1979a? Connolly, 1980). With a method which
does not capture or provide readily located victims, application
of the more stringent criteria of Griffiths et al. (1978) is
needed in order to draw any inferences concerning the usefulness

of the method.
D. Mechanical Control Methods
The Cain Commission (Cain et al., 1972), -the conservation
groups petitioning the EPA in 1972 (Natural Resources Defense
Council et al., 1971) and, ultimately, the Administrator of SPA
(Ruckelshaus, 1972) were in agreement that methods available as
alternatives to the use of toxicants were sufficient to control
predator damage to livestock. The Administrator noted:
For the maintenance of predator control programs, es-
pecially in the sheep industry, effective non-chemical
alternatives exist, including denning, shooting, and
trapping, methods that have long been available and
effective, though more costly than poisons. (Ruckels-
haus, 1972, p. 5720)
Table V-l summarizes the assessments of these methods which
appeared in the major support documents reviewed by the Agency
in 1972. Studies of these methods have indicated some support
for these conclusions as well as some contradictions and some
areas which merit further investigation.
1. Aerial Gunning
Cain et ad. (1972) recommended that this practice be limited
to "authorized biologists of the appropriate Federal and State
Agencies" (p.9). The commission did not believe that the method
would be used very often. Since 1972, however, aerial shooting
has accounted for a substantial proportion of the total taiown
coyote take by the USDI Animal Damage Control (ADC) Program. In
1976, for example, more than one third of the coyotes known to
have been taken in the ADC program were shot from aircraft (USDI,
1978). The method is costly on a per-hour basis with costs
increaing sharply in recent years (Glosser, 1981; Treat, 1981;

Koch, 1981) Although the take/man year of effort tends to be high
on the average (USDI, 1978), the productivity of aerial gunning
of coyotes can vary considerably from region to region and season
to season (Anderson et al., 1974a, b). In brushy areas, for
example, locating coyotes from the air can be difficult.
Aerial gunning is practiced in areas where livestock damage
is occurring. Diphacinone residues (from the DPN toxic collar
trials) were found in several coyotes shot from aircraft, indica-
ting that "offending" coyotes are taken by this method (Connolly,
et al., 1980).
2.	Ground Shooting
Animals shot from the ground may be lured into open areas by
use of calls or dogs or they may simply be hunted. Although the
USDI's (1978) assessments of ground shooting are at variance
with those of Cain et al. (1972-see Table V-l), there has been
very little new information developed on ground shooting since
the toxicant ban.
3.	Denning
In this method, coyote dens and their occupants are destroyed.
The method can only be used during the spring months when dens
are used. The theory behind denning is that not only are coyote
pups killed, but also the need for the adults to obtain large
amounts of food is removed (Young and Dobyns, 1945). As a result,
a pair not taken may decrease its rate of predation or stop
taking livestock entirely when its pups are taken through denning
(USDI, 1978). Although Cain et al. (1972) favored this method
as an effective troubleshooting tool, denning is no longer prac-
ticed by ADC personnel following a directive from the Secretary
of the Interior (Andrus, 1979).
4.	Traps
The steel leg-hold trap is the most widely used coyote trap.

It usually is not lethal to its victim/ although target animals
captured are generally destroyed. Non-target species are fre-
quently caught in these traps, but selectivity for target species
reportedly can be improved by selecting the proper type of set
and scent (USDI, 1978; Boddicker, 1981). Efficient, selective use
of steel leg-hold traps requires special skills and experience.
Reports of total takes of animals (e.g., USDI, 1978, p.79) do
not accurately reflect the selectivity of the method since dif-
ferent species may be targeted in different trapping situations.
Recent work by USDI personnel has been directed toward improving
humaneness and selectivity by using tranquilizer tabs and pan
tension devices.
Traps are useful in corrective situations, but the time and
labor needed for their deployment renders them inefficient in
prophylactic control operations. Cain et al. (1972) and the
conservation groups (Natural Resources Defense Council, 1971)
believed that extension programs could be used West-wide for
training livestock producers in the proper use of steel leg-hold
traps, thus placing much of the manpower burden for predator
control on the producer himself. Such a system has been reported
to be successful in Kansas (Henderson, 1972; Robel, 1981). Bod-
dicker (1981) doubts that this approach can be successfully ap-
plied as the only predator control system in the more mountainous
5. Neck-Snares
Neck snares were not discussed in the support documents used
by EPA in 1972. Snares are loops of wire used to choke target
animals as they pass through restricted areas such as habitual
places of travel through or under fencing. Non-target animals
are sometimes taken by snares. Careful selection of sites and
knowledge of the animals in the area can improve selectivity
(USDI, 1978). The effectiveness and selectivity of neck snares
in Texas has recently been reported by Guthery and Seasom (1973) .
These authors believe that snares can be effective in prophylactic

control programs in areas in which woven-wire fencing is used
extensively. In other areas, snares may be "poor" in population
reduction efficiency and in cost per coyote taken (OSDI, 1978) .
S. Livestock Husbandry Practices
In recommending expanded extension programs to enable sheep
producers to resolve their problems with predators, Cain et al.
(1972) suggested that these programs should encourage "the use
of better husbandry and management practices" (p. Ill) as well
as instructing producers in mechanical control techniques.
Since the toxicant ban, there has been considerable public debate
over management practices (for examples, the 1978 Animal Damage
Control Policy Study hearings and the 1981 EPA information gath-
ering hearings)... Opponents of the use of 1080 for predator
control generally urge livestock producers to practice better
husbandry (cf. Atkins, 1981; Hoff, 1981). Proponents of reintro-
duction of 1080 argue that producers are now practicing all of
the known antipredation husbandry techniques suitable to their
types of operations which are economically and logistically
feasible (cf. Helle, 1981; Hibbard,1981). Support for both
sides' contentions and, consequently, illustration of the bind
in which the livestock producer may find himself, is found in
the work of Faulkner and Tigner (1977). These researchers found
that the practice of shed lambing consistently increased the
number of lambs per 100 ewes that survived until docking over
the docking percent obtained for lambs born on the open range to
ewes from the same flocks. For the areas and types of operations
studied, however, shed lambing for entire bands of sheep is
prohibitively expensive (Faulkner and Tigner, 1977).
Nass (1980a, b) has recently reviewed the various husbandry
practices thought to reduce predation (Table V-2) and has identi-
fied the types of operations on which these practices could be
used .(Table V-3). Most of these practices are fairly traditional
in American sheep raising although new data are being gathered
on their utilities. The use of dogs to guard sheep flocks is

Husbandry practice
Confinement raising of sheep
Confinement raising of lambs
Confinement at night
Selective use of pastures
Check sheep daily
Additional herders
Barriers in large pastures
Close herding
Shed lambing
Carrion disposal
Keep flock healthy
Change lambing dates
Shorten lambing period
Dnproved fencing
Table V-2
Husbandry Practice Tradeoffs
Hay reduce predation
Problems easily seen
May reduce predation
Gain weight faster
May reduce predation
Problems easily seen
May reduce predation
May reduce predaticn
Problems easily seen
May reduce peedation
May reduce predation
Problems easily seen
May reduce predation
Prcblems easily seen
Increased lamb survival
Predaticn reduced
May reduce predaticn
More lambs marketed
Possible less predation
Lambs absent in critical
Seduce snail lands exposure
Ctiifocn marketing improved
May reduce predation
Better grazing
Limits predator access
More disease problems
Higher feed costs
More disease problems
Higher feed ccsts
Limits grazing time
Mare labor involved
May waste forage
Possible reduction of sheep
Mare labor involved
Overhead increased
Higher ccsts
Higher ccsts
Soil compaction and ercsion
Vegetative compaction
Disease increased
More labor required
Initial cost high
Mare labor involved
May increase feed ccsts
More labor required
May not fit labor pattern
Forage may not be available
Intensive labor requirements
Sheds may be too snail
Initial costs higher
Higher costs to maintain

fable V-2 (Continued)
Husbandry Practice Tradeoffs
Husbandry Practice
Truck sheep instead of trail
May reduce predation
Less snail lamb mortality
Costs are high
Extra Seed may be required
P»T*ted Predation Reduction Practices
Lighted corrals
Bells on sheep
Guarding dogs present
Electric fencing
Hay reduce predation
Easier to check on sheep
May reduce predation
May reduce coyote
Seduces predation
Keeps stock £raa straying
Increased costs
Cost of bells and collars
Labor increased
Questionable efficacy
May increase dog predation
Costs increased
Casts higher
High costs to maintain
Feat N&ss, 1980b.

Tafale 7-3
A List of Livestock Husbandry Practices that May Beducs Predatlon
applicable Application
m-wahandry Practice
Cbnfinenent raising of sheep
Confinement raising of lambs
Confinement at night
Selective use of pastures
Check sheep daily
additional herders
Herders in large pastures
Close herding
Shed lanbing
Carrion disposal
Keep flock healthy
Change lanbing dates
Shorten lanbing period
Imgcoved facing
Truck sheep instead of trail
Belated predator Reduction practices
Lighted corrals
Bells on sheep
Guarding dogs present
electric fencing .
Facn Flock
Large Pasture
Fran ttass, 1980b.

relatively new in this country. This practice is reviewed brief-
ly below.
1. Guard Dogs
Interest in the use o£ dogs to protect sheep from predation
has increased greatly in the ?.S. in recent years. Spurred by
government and university research, this approach has apparently
been effective in reducing predation on certain operations (Cop-
pinger, 1980; Green and Woodruff, 1980). Some opponents to the
reintroduction of 1080 as a predacide favor the use of guard
dogs (e.g. Atkins, 1981; Stevens 1981).
The breeds of dogs which researchers, ranchers and faraers
have evaluated as livestock protectors are types which are used
in Europe and Asia to protect sheep. These breeds include Komon-
dor, Kuvasz, Great Pyrenees, Anatolian Shepherd, ovcharka, Marem-
ma, Sar Planinetz, and Karabash (Coppinger, 1980; Green et al.,
1980)« The breeds which can be used effectively as livestock
guard dogs differ behaviorally from the sheep herding breeds.
While herding, dogs frequently direct elements of hunting behavior
sequences (e.g., stalking) toward the sheep, guard dogs* re-
sponses to their charges tend to be more filial (Coppinger, 1980)
Socialization of dogs to sheep (or other livestock) is an
essential element in the development of guard dogs. A properly
socialized guard dog will exhibit the following behaviors neces-
sary for livestock protection: "nonaggression toward the sheep,
attentiveness to the sheep (* following instinct'), and defense
of the sheep" (Coppinger, 1980). Proper socialization includes
exposure to livestock and livestock operations at an early age
and some general training in obedience and for the elimination
of undesirable behaviors such as the harassing of livestock
(Green and Woodruff, 1980). Using dogs that were probably too
old for complete socialization toward sheep, Linhart et al.
(1979) reported some harassment of sheep by guard dogs. Never-
theless, coyote predation on sheep appeared to have been sup*
pressed in this study by the presence of dogs (xomondorok).

Even when conditions are optimal for socialization and train-
ing, not all individuals of the sheep guarding breeds actually
become effective livestock protectors (Green and Woodruff, 1980).
The use of dogs may not be appropriate for all situations or for
all producers (Green and Woodruff, 1980) . For example, at the
information gathering hearings, Popoulas (1981) and Howard (1981a)
reported lack of success in using guard dogs. Green et al.
(1980) have recently summarized some of the economic factors
involved with the use of guard dogs, an historically old practice
that has only recently been promoted in this country. The pro-
ducer must spend a considerable sum of money to acquire and feed
a dog and this dog must be conditioned to the livestock operation.
Whether the dog "pays for itself" is determined by the amount of
relief from predation which can be attributed to the presence of
the dog.*
F. Effects of Coyote Control on Livestock Losses
Beyond the question of whether an individual method kills,
repels or captures coyotes lies the question of the value of
coyote control in reducing losses of livestock. The primary
reference documents used by the Agency in 1972 (Natural Resources
Defense Council, 1972; Leopold, 1964; Cain et al., 1972) and the
Administrator himself questioned the benefits derived from the
use of predacides (RucJoelshaus, 1972).
Since 1972, the percents of sheep and goats taken by coyotes
and/or other predators have been estimated in many publications.
While there is much variation among studies in loss estimates,
in methods of estimation, and in other relevant areas, lower
estimates are usually reported for areas in which predator control
has been practiced than where no control measures were used (cf.
Section IV; OSDI, 1973; Nass, 1980c).
* Dogs working in areas where 1080 baits were placed would
probably be killed if they consumed baits. Any program for re-
introduction of 1080 baits should provide for protection of
herd and guard dogs.

The exact relationship between coyote control and the loss of
livestock to coyotes is not defined. The Cain Commission opposed
the idea of obtaining relief from livestock damage through cam-
paigns designed to achieve general suppression of coyote popula-
tions. The Cain Commission stated that localized loss problems
could be resolved by selective removal of the individual ("of-
fending") animals responsible for the damage. Proponents of
predator control have stated that individuals engaged in predator
control must have a large variety of control methods at their
disposal to meet the demands of the variety of circumstances
which arise (Grieb, 1981; Beck, 1981) . Cne such circumstance is
said to be the situation in which local coyote population reduc-
tion or extermination is the most efficient (or best, or only)
way to stop predation on livestock. Another problem situation is
said to arise when coyote predation cannot be stopped by the use
of the control techniques now available.*
Despite the controversy stirred by the topic of coyote con-
trol , there are several areas of tacit agreement among nearly
all parties. The first is that in the absence of coyotes, there
are no livestock lost to coyotes. Although some, including Cain
et al. (1972), have argued that many lambs lost to coyotes would
have been lost to other causes, the seemingly trivial point that
no coyotes present means no loss to coyotes gives rise to the
second area of tacit agreement: that coyote predation on live-
stock is in some way related to coyote density. The way(s) in
which livestock loss may be related to coyote density are not
known and have not been subject to a great deal of research or
theoretical modeling. The third area of tacit agreement is that
by killing, repelling or mitigating the offending coyotes, one
can stop predation by coyotes. The unanswered questions in this
area involve the most efficient way(s) of controlling offending
* Many of thosewho believe that currently available methods
are not sufficiently effective believe that effective control of
coyote predation on livestock could be achieved by the addition
of 1080 to the control agent's arsenal (e.g., Bowns, 1981; Beck,
1981; Grieb, 1981; Meike, 1981).

coyotes and the changes in the behavior of surviving coyotes
which are precipitated by removal of offending (or other) coy-
At the information gathering hearings seme opponents of
predator control (e.g., Morris, 1981; Ryden, 1981; Strojny, 1981)
have argued that disturbing coyote population through predator
control efforts actually increases predation through a variety
of effects including stimulation of coyote reproduction (both
'in percent of females reproducing and in average litter size per
whelping bitch), and stimulation of immigration by opportunistic
coyotes which would be more likely than the former residents to
select livestock as prey. Where toxic baits are used, Ryden
(1981) argued that coyotes with a tendency to feed on carrion
would selectively be removed from the local population, while
the more strictly predacious conspecifics would survive. While
there is no clearcut evidence to support many of these conten-
tions, Knowlton (1972) has reported greater average numbers of
uterine swellings in female coyotes captured in areas where
predator control efforts were intense than in females from areas
not subjected to extensive control programs. That high, stable
populations of coyotes can exist in the same area as livestock
operations without, significant amounts of losses may be ques-
tioned since these carnivores would be required to exist on
available supplies of natural live prey and carrion, both of
which are subject to seasonal and other cyclical fluctuations in
availability. As support for the contention that such coexis-
tence is possible, Ryden (1981) cited lower levels of sheep and
lamb losses in Kansas than in Wyoming despite the higher coyote
index in Kansas. The validity of such a comparison may be ques-
tioned, however, due to difference between the two states in
climate, topography, predominant types of livestock operations
and habits of natural prey (cf. Boddicker's [1981] distinction
between conditions in Kansas and Colorado).
While there is some agreement that not all coyotes living
close to ranch or farm operations will prey upon livestock, the
proportion of coyotes living under such conditions that are (or

may become) livestock predators is not known.* The Cain Commis-
sion and many of the representatives of conservation groups
testifying at the information gathering hearings (cf., Armentrout,
1981; Wentz, 1981; Reed, 1981} have implied that only a small
proportion of local populations o£ coyotes prey upon livestock.
Robel (1981) has recently stated that removal of one to three
coyotes resolved predation problems for about 40 percent of the
cooperators in a recent study.
Although some of the field studies (Connolly, 1980; Connolly
at al., 1978; USUI/ 1979b) conducted with the toxic collar appear
to support the notion that removing a few offending animals can
resolve coyote predation problems, these data are subject to
other interpretations. The observation that losses stopped
after a few collars were punctured is also consistent with the
theory that a few immigrants had been removed from a previously
coyote-free area.**
When predation stops after collars are used in conjunction
with other methods, one cannot be sure whether the collar con-
tributed by taking the "true" offending animals or by serving as
the "last straw" in a general population reduction effort. For
example# the toxic collar has been tested, for two years near
Meridian, Texas, on several goat ranches which are managed by
one individual. Prom July 1979 through June 1981, thirty-six
punctures of collars have been confirmed or appear to have oc-
curred (Howard, 1981b). During that same period, 266 coyotes
* The percent of coyotes that are "offenders" is probably not a
stable value. The proportion could vary considerably £rom one
situation to another. Althoff and Gipson (1981) recently reported
that 3 of 19 radio collared coyotes (from 2 of 8 "families") known
to range within 5 tan of two Nebraska turkey operations were known
to have killed turkeys. However, known home ranges for these
animals showed that only one family's range overlapped the produc-
tion sites extensively. Untagged coyotes were also believed to be
involved in predation, leaving uncertain the estimate of the per-
cent of offenders among the local coyote population.
** The same interpretation could be applied to explain the ob-
servations reported by Robel (1981; Robel et al., 1981) unless
it were known that coyotes remained in the damage area after "of-
fending animal" control had stopped livestock losses.

are known to have been taken by other control methods used on or
near the properties where the livestock (principally angora
goats) have been run. While this effort has led to significant
reduction in rates of loss to coyotes, the data do not indicate
the relative contribution of the various methods to the reduction
of predation. Since the coyotes taken by the collar were clearly
predators on livestock/ at least 12 percent of the total number
taken were offending animals. It is also possible, however, that
all coyotes taken had preyed upon livestock at least once.
If it can be shown that nearly all coyotes living in proximi-
ty to livestock operations are (or can be expected to be) signif-
icant predators on livestock, predator population supression may
provide an efficient way of reducing predation. A simulation
model developed by Connolly and Longhurst (1975) suggests that
coyote extermination over large geographical areas would require
massive effort,* given the apparent Capacity of the species to
intensify its reproductive effort in response to ccfttrol programs.
If future studies show that offending coyotes constitute a small
(and relatively constant) portion of coyote populations, prophy-
lactic control programs would not be efficient (or necessary)
for resolving predator problems. Since two of the proposed uses
of 1080 (bait stations and single lethal dose baits) are primari-
ly sought for purposes of population suppression, the usefulness
of such approaches in reducing'predation depends directly upon
the proportion of the local coyote population which preys upon
The timing of predator control may also affect its efficacy
in reducing livestock loss. The notion that control is most
effective when practiced just before or during the coyote's
reproductive season dates back at least as far as Robinson's
(1948) report and has recently been reiterated by Dorrance (1980)
* The cost of such an effort would depend upon the methods
used to achieve control. The use of toxic baits is reportedly
much more economical than mechanical control methods (cf.,
Boddicker, 1931; Bourret,1981)

-8 6-
who studied the use of toxicants by livestock producers in Al-
berta.* It is possible that the loss o£ 1030 bait stations and,
more recently, denning from the ADC program's arsenal have hin-
dered federal efforts at disrupting coyote reproduction in high
loss areas. If this is true, the reported higher livestock loss
levels and higher takes of coyotes in the post 1972 period may
not be irreconcilable.** Inappropriately timed efforts may have
led to a need to take more coyotes.
The period (1972 to the present) since the cancellation of
uses of 1080 as a predacide has seen an unprecedented research
effort directed toward various aspects of coyote predation on
livestock. Host of"this research, however, has either documented
the existence of predation problems or demonstrated the utilities
and limitations of control and management methods which do not
involve the use of 1080. Much of this new information has been
reviewed in this section. The toxic collar is the only proposed
use of 1080 for which a significant amount of data pertinent to
the prevention of coyote predation on livestock have been gener-
ated (e.g., Connolly, 1980; tJSDI, 1979b).
Prior to the 1972 cancellations of predator control toxi-
cants, little effort was expended to document the effectiveness
of predacides'through careful research (Balser, 1974). To date,
there have been no sound research data published which demon-
strate the effectiveness of 1080 single dose drop baits in con-
trolling coyotes in the U.S. The effectiveness data (Robinson,
1948), along with an attempt to derive effectiveness data from
total coyote takes and manpower data (Wagner, 1972), for the 1080
bait station have been reviewed in this section. On the basis of
* Dorrance suggested that strychnine baits could be used most
efficiently in Alberta if they were applied only in March and
April. These late applications would, according to Dorrance, dis-
rupt the reproductive process while at the same time taking advan-
tage of winter mortality in coyotes and exposing non-target animals
to toxic baits for shorter periods of time.
** The proportion of the coyotes actually killed in the ADC
program by toxicants used prior to 1972 that was included in the
program's reported totals of animals taken is not known, however.

what little evidence is available for these methods/ it is pos-
sible only to -speculate concerning the effects that their intro-
duction would have on the level of coyote predation on domestic
The results of the toxic collar research indicate that the
method can be used successfully by biologists as a corrective
tool to remove coyotes that attack sheep or goats in the throat
region. In some situations, the collar may do the complete job
of eliminating predation, but other methods of control are often
needed. The method would probably be of little value on range
operations or in other situations in which it is difficult to
direct predation. Addition of this method to the predator con-
trol arsenal could help some family farm and/or fenced pasture
operators to stay in the sheep or goat business. Whether this
would occur would depend upon whether the use were registered
and whether it would prove to be possible to set up efficient
systems for training ranchers to use collars effectively and
for keeping track of the fate of individual 1080 collars.
From the testimony at the information gathering hearings, it
is apparent that many livestock producers perceive currently
available mechanisms for controlling coyotes to be insufficient
or impractical for all damage situations. Although other influ-
ences have contributed/ it is evident that animal losses to
predators is seen as a major factor in the decline of the sheep
industry in the O.S. To the extent that any new approaches to
the coyote problem would be sufficiently effective to enable
individual producers to stay in business, the methods would help
industry. Inexpensive, efficient and safe means of reducing
damage are needed to augment (or partially replace) control
methods currently used. Reintroduction of 1080 as a predacide
might fill this need, but the efficacy of the proposed uses has
not been fully established (see Section VI for discussions of
safety of proposed uses).

A. Hazards to Wildlife
1. Information on Non-Target Hazards of 108Q Used
in 1972 Decision
One of the primary reasons for suspending and- cancelling
the use of 1080 to control predators was the hazards it posed to
non-target wildlife. The Agency found that 1080 was extremely
toxic to all species and that indiscriminate baiting with 1080
over wide remote areas posed two hazards to non-target animals:
1) primary poisonings of non-target animals that feed on baits
placed for target species and 2) secondary poisonings of non-
target animals that scavenge remains of poisoned animals. The
order cancelling the use of 1080 to control predators indicated
that while the impacts on non-target species from the use of
1080 to control predators were for the most part undocumented*
the available evidence may well have underestimated the true
damage. The order further stated:
It is appropriate to take administrative notice
of the fact that isolated accidents involving wildlife
are not apt to be reported. Isolated, even if routine
and numerous instances of secondary animal poisoning
would not have the visibility of a wildlife kill nor is
there apt to be an observer present as in the case of
human mishap. The administrative process need not be
blind to these realities.
The order maintained that the use of 1080 in large bait stations
posed an imminent hazard especially to endangered species and
stated that the death of even one animal which belongs to an
endangered species is an irreparable loss because it renders
such species closer to extinction.
As noted earlier, the Agency relied heavily on information
presented in "The Cain Report," The Natural Resources Defense

Council petition to ban 1080/ and "The Leopold Report."
Cain et al. (1972) briefly reviewed the available toxicity
data on 1080. Table VI-1 summarizes toxicity data from sources
cited in Cain et al. (1972). The Cain Committee indicated that
in comparison to other toxicants used in predator control/ 1080
was more toxic to canids than to most other species. In other
words, when species are compared* on the basis of the amount of
1080 required to kill an average animal (expressed as ng of 1080/
lag of body weight), canids are more sensitive to 1080 poisoning.
This characteristic is called differential toxicity.
Proponents of 1080 argue that the differential toxicity
would allow users to minimize hazards to non-target wildlife by
controlling the concentration o£ 1080 in a bait station. "The
Cain Report" cites Martin and Atzert as examples of this line of
Martin (1971) stated:
Although sodium monofluoroacetate is generally highly
toxic, there is sufficient range of sensitivity between
species to allow a degree of selectivity through for-
mulating practices. Since the compound is highly
soluble in water, it is possible to inject an aqueous
solution into large pieces of meat which then may be
securely fastened to the ground. When treated ac-
cording to standardized Bureau directions, a coyote
need eat only 1.4 ounce of treated meat to receive an
LD50. In contrast, a golden eagle must eat about
12 ounces, a great homed owl about 1 pound, a black
vulture over 2 pounds, a bear from 4 to 8 pounds, and
a human must eat from 3 to 8 pounds to obtain a lethal
dose. This characteristic makes sodium monofluoroace-
tate unique for use in meat bait stations that are
placed in remote locations during the fall and winter
months to control coyote populations.
Atzert (1971) argued:
The golden eagle, an animal that normally consumes the

Table 71-1. LDgo's of sodium monofluoroacetate.
of Admin-
Man	0.7-2.1	Estimated	Oral	1,2
Rhesus monJcey
mulatta)	4.0	1.7.	3
Spider monkey
geoffryi)	13.0	l.v.	3
Opossum (Didelpbia
marsupialis)	60.0	Oral	9
adults (P)	0.393	0.247-0.525	Oral	4
(34—P)	0.221	0.149-0.327	Oral	4
Goat	0.6	X.M.	3
Horse (M-F)	0.35-0.55	Oral	5
Mule (M-F)	0.22-0.44	Oral	5
Mule Deer
h. hemionus)
M-F	0.30-1.00	Oral	5
Sheep (M-F)	0.25-0.50	Oral	6
adult	<1.0	Oral	3
young	0.4	Oral	3
3ear (Urus
sp.)	0.5-1.0	Oral	7
Bobcat (Lynx
rufus baileyi)	<0.56	I.P.	8
Domestic cat	0.20	X.V.	3
Coyote (Canis
nebracensis)	0.10	X.v.	8
Grey Fox (Urocyon
scotti)	<0.3	X.P.	8

Table VI-1 (continued)



o£ Admin-

Badger (Tasidea



Domestic £erret



Oral (S.T.)
Marten (Martes


Mink. (Mustela



Ground Sauirrelss


(Citellus c.


Fisher* s


beecheyi fisheri)

Pocket Gophers:





Tuza (Geomys


Kangaroo Rats:


(Oipodomys s.



(Oipodcmys m.



Table 71-1 (continued)



of Admin-




norvegicus) M



(Rattus rattus


Black (Rattus

rattus sp.)





Norway-wild (Rattus



wood (Neotoma

a. albigula)

Wood (Neotoma



Deer mouse

(Peromyscus so.)

Bouse mouse

(Mus musculus)

Miscellaneous sop:

Meadow vole









Prairie Dog



Oral (S.T.)

Table VI-1 (continued)



of Admin-



jack rabbit



European Rabbit





Domestic pigeon


livia) (M-P)
Mourning Dove


macroura) (H-F)
3 .35-14.6

Oral (S.T.)

Mallard (Anas p.


adult (M)

Oral (S.T.)
adult (?)

Oral (S.T.)

(Anas acuta


adult (M)

Oral (S.T.)
adult (F)

Oral (S.T.)


Chukar (Alectoris

graeca) (M-F)
Gambels quail

(Lophortyx gambeli)

Japanese Quail



japonica) (M)




colchicus) (U)


gallopavo) (F)

Table VT-I (continued)



of Admin-


Brewer* s




English Sparrow


domesticus) (M)
Magpie (Pica p.


Raptors and


Golden eagle





American rough-

legged hawk

(3uteo lagopus

— 10 .0***

Ferruginous rough-

legged hawk

(Buteb regalis)

Marsh hawk

(Circus cyaneus

—10 .0***

Great Homed Cwl

(Bubo virginianus


Black vulture



Turkey vulture

(Cathartes aura)

Oral (S.T.)

Bull Frog (Rana

catesbeiana) (M)
25 .6-115
Leopard Frog (Rana


South- African

Clawed toad

(Xenopis laevis)


Footnotes to Table VI-1.
1.	Kaye (1970)
2.	Arena (1970)
3.	Chenoweth (1949)
4.	Bobison. (1970)
5.	Tucker and Crab tree (1970)
6-.	Jensen et al. (1948)
7.	Robinson (1953)
8.	Ward and Spencer (1947)
9.	Denver Wildlife Research Center (Unpublished)
10.	Lasarus (1956)
*	Where confidence limits are not provided the figure is
assumed to be an observed non-statistical estimate.
**	Research has shown much variation between strains of
laboratory rodents (Chenoweth, 1949).
***	Vomiting characteristic and early symptom.
»	Male
F	Female
1.7.	Intravenous
I.M.	Intramuscular
I.?.	Intraperitoneal
S.T.	Stomach Tube
S.C.	Subcutaneous
<	Less than
>	Greater than
—•	Approximately
From Atzert (1971)

viscera before other portions of its food, demonstrates
the low hazard of acute poisoning via secondary sources.
To obtain an LD50 (1.25-5.00 mg/kg) of sodium mono-
fluoroacetate from a secondary source such as coyotes,
a 7-pound golden eagle must consume the viscera of
from 7 to 30 coyotes killed by sodium aonofluoroacetate,
assuming the coyotes ingest LD50	og/Iog) and do
not excrete, detoxify, or regurgitate any of the toxi-
cant, and that as in rats approximately 40 percent of
the toxicant is present in the viscera at death. The
viscera of coyotes account for approximately 20 to 25
percent of their live weight or 6 or 7 pounds. A golden
eagle's daily consumption of food equals approximately
30 percent of its live weight, or 2 pounds (Denver
Wildlife Research Center, unpublished data). As noted
previously, animals can metabolize and/or excrete con-
tinued small doses of sodium monofluoroacetate without
Both the Leopold and Cain committees agreed with these
arguments, at least in theory. "The Leopold Report" concluded
"that when properly applied, according to regulations, 1080
stations do an effective and humane job of controlling coyotes
and have very little damaging effect on other wildlife." "The
Cain Report" expresses this same thought, stating:
If this [1030] and other toxieants were consistently
applied under field conditions with the meticulous
care specified in the- operation manual, it is quite
possible that a major portion of undesirable side
effects could be avoided.
However, "The Cain Report" contended either that 1080 was
not being applied carefully or that the persuasive assumptions
made in laboratory tests did not apply. To support its conten-
tion "The Cain Report" referred to evidence indicating that

non-target species were being killed. In particular, the report
cited testimony by Alexander Sprunt IV o£ the National Audubon
Society before the Senate Appropriation Subcommittee on Agricul-
ture in 1971, that a bald eagle, a California condor, and 11
golden eagles had been killed by 1080. (See Table VI-2).
"The Cain Report also referenced Robinson (1943) who reported
the results o£ tests evaluating both 1080 and thallium in large
bait stations. "The Cain Report" cited the following incidents
reported by Robinson (1943):
Both thallium and 1030 are relatively slow in their
toxic action, with the result that the creatures that
succumb after feeding on the station are scattered
over such wide areas that complete counts can
never be made. Stockmen, sheepherders, service
hunters and others working in the experimental area
following the use of the stations have reported on
the creatures found dead, presumably poisoned; the
combined reports from these sources list the following:
383 coyotes, 3 bobcats, 37 dogs, 1 domestic cat, 2
badgers, 4 weasels, 8 eagles, 7 magpies, 4 hawks, and
2 ground squirrels. Some of these men were careful
observers, but undoubtedly the majority were interested
primarily in predators, and therefore the compilation
may be considered as emphasizing the coyote.
The search of another baited area by crews looking
specifically for all species of dead animals turned
up 61 coyotes, 4 badgers, 1 mink, 28 magpies,
. 2 hawks, 3 eagles, 9 deer mice, and 6 ground squirrels.
The best indication of the extent of secondary
feeding was furnished when crews were used to search
station sites for victims or their scattered remains.
During 39 man-days of hunting, 18 coyotes, 20 magpies,
2 golden eagles, and 2 hawks were located, of these, 9

lies a
Tatole 71-2
Denver Wildlife Research Center-Denver, Colorado
1080 and Strychnine in Reports
Scuth Dakota
Sooth Dakota
Scuth Dakota
South Dakota
South Dakota
South Dakota
Scuth Dakota
Saaple Description
CaemictL Found
California Condor, 1080 - 0.75 mg.
stomach lining, crop
contents, heart tissue
California Condor,
stnoach contents
Golden Eagle, heart,
liver and stomach
Golden Eagle, heart,
liver and stomach
Pa 7r? Eagle,
No str*±nine
1080 - 0.13 og.
Ho atrydjiine
1080 - 0.20 tag.
1080 - 0.24 mg.
Uo stejehnine
Golden Eagle, viscera 1080 - 0.55 mg.
(liver and G.I. tract)
Golden Eagle, viscera 1080 >0.21
(liver and G.Z. tract)
Golden Eagle, viscera 1080 - 0.29 mg.
(liver and G.I. tract)
Golden Eagle heart,
liver and G.I. tract
Golden Eagle stomach
Golden Eagle stonach
Golden Eagle stsnach
Golden Eagle stomach
Golden Eagle viscera
Golden Eagle viscera
1080 - 0.31 ag.
1080 - not found
1080 - 0.33 tog.
1080 - 0.02 mg.
1080 - 0.05 mg.
1080 - not found
1080 - net found

Table VI-2 (Continued)
Ingoiea Ntmhftg
Centres Wildlife Besearch Center-Denror, Colocado
1080 and SUiyiaxine in Blasts
South Datota
South Dakota
South Dakota
Sample Cescricr.icn
Golden Eagle, head
CyaldL Staund
SBCST - present
Golden Eagle, viscera Strychnine - Sound
1080 - not found
Golden. Eagle, viscera Strychnine - found
1080 - not fcurd
Golden Eagle, asm
Strychnine - ncne
Ifl80 - 0.36 ag«
Golden Eagle, viscera Strychnine - ncne
1080 - 0.1 ng.
Date Bscaived
Stan Cain et al., 1972

o£ the coyotes had been eaten in typical eagle fashion,
and at least 8 of.the magpies either were completely
eaten, with only scattered feathers as evidence* or
the remains were so dispersed as to suggest scavenger
"The Cain- Report" further indicated that Robinson (1948)
concluded that eagles were the most likely non-target to he
killed by secondary poisoning because of their feeding pattern
(eating the viscera of poisoned coyotes), but surveys conducted
during periods of large bait station use showed no significant
reduction in the population of these birds.
Both "The Leopold Report" and the Natural Resources Defense
Council petition to ban 1080 contained information concerning
misuse of 1030. "The Leopold Report," although giving no sup-
porting information, indicated that in many instances, regula-
tions were not followed: 1080 stations were placed much closer
together than they should be; excessive amounts of poison were
used; and the poisoned bait was not always picked up in the
spring. The report indicated that, under these circumstances,
considerable damage could occur to wildlife. The Natural
Resources Defense Council petition referred to a study conducted
by Dr. Alfred Etter, who found that in one region of Colorado,
63 townships, or one-third of the baited townships there, con-
tained three or more baits. In 18 of those, from five to 15
stations had been approved. The study also,*J.n<3icated that there
was a ten fold variation in station weights, poison was not
~ Questions have been raised about the use of the field
mortality information associated with large bait stations cited
in "The Cain Report" to evaluate the hazards of 1080. The in-
cidents cited in "The Cain Report" were from Robinson (1948),
which reported the results of tests evaluating the use of large
bait stations over a nine year period. The first seven years of
the study dealt only with thallium, while the. last two years
also included 1080. Because of the manner in which animal deaths
were reported, it cannot be determined whether 1080, thallium,
or both were the cause of death. Therefore, this part of
Robinson's work is of limited value in assessing the hazards of
1080 to non-target species.

injected uniformly in baits, and that over half of the users
exceeded the norm, some putting more than twice the standard .
dosage in their baits (Etter 1968/ 1969 in the Natural Resources
Defense Council petition to ban 1080).
"The Cain Report" emphasized that evaluation of impacts
must be on potential hazard, since data on which to base firm
judgments, with the possible exception of data indicating hazards
to endangered species, are scant. The report further added that
individual animals of a wide variety of species have been de-
stroyed by 1080 despite the greater toxicity of 1080 to canids.
But, the report pointed out, that the death of some non-target
animals does not necessarily result in a material reduction of
the population of the species, unless the species is endangered.
"The Cain Report" then noted that 1080 was thought to have caused
the death of a Sierra del Nido (Mexico) grizzly bear and two
California condors, both endangered species.
Based on this information, the Administrator found that
there was evidence that a certain number of non-target animals
were being adversely affected by 1080 products, particularly
carrion eating birds and mammals. While EPA recognized the
uncertainties about how various animal populations were being
affacted from poisoning of individuals, SPA found that the avail-
able evidence showed that 1080 had contributed to the death of
endangered species. Also, the Agency believed that in many
instances use directions were not followed, increasing potential
damage. The Administrator concluded that the predator use of
1080 presented an imminent hazard to non-target wildlife, in-
cluding some endangered species, and that suspension and cancel-
lation was warranted.
2. Mew Information Since 1972 on 1080 - General
Since the 1972 ban on predacidal use of 1080, several points
have come to the attention of the Agency in relation to evalu-
ating potential hazards of using 1080 to control predators.
Arguments presented either supporting the safety of 1080 or

disputing the safety of 1080 when used to control predators have
in several instances relied on chemical analysis results and/or
published toxicity data on 1080 (Ketron, 1979? Audubon, 1971;
Morton, 1971; At2ert, 1972; Connolly, 1980). In the Agency's reg-
ulatory review of applications to use 1080, it has become apparent
that neither the chemical residue analyses nor the toxicity data
are well defined.
The U.S. Fish and Wildlife Service's Denver Wildlife Research
Center {DWRC) has reviewed the four basic analytical methods
for 1080 (Okumo in Connolly, 1980). These methods were aconitase
enzyme inhibition by citrate formation, colorimetry of fluoroacetic
acid, measurement of total fluorine by colorimetry or ion-selective
electrodes, and gas chromatography. This review indicates that
each of the methods has one or more of the following shortcomings:
the method is not quantitative; the method is not specific to
1080; and/or the method's reliability is not known. Okumo concluded
that currently the best method for detecting the presence of low
levels of fluoroacetate (1080) in animal tissues appears to be
the Okumo and Meeker (1980) method using gas-liquid chromato-
graphic determination.
The Okumo and Meeker method, however, appears to. be less
than'completely reliable (Connolly, 1980). In tests to evaluate
its reliability, several blind duplicates and one known negative
tissue were submitted for analysis. Of three pairs of blind
duplicates of muscle samples that were analyzed, two yielded
values that agreed closely. The third pair of values were con-
sistent in that both showed high 1080 content, but one result
was more than double the other. Analysis of the sample that was
known to contain no 1080 indicated that it contained trace amounts
of 1080. Also, false negatives appeared. Of three different
coyotes that were known to have died of 1080 poisoning, none was
reported to contain 1080.
This latter result could be explained by the mode of action
• %
of compound 1080. Connolly (1980) speculated that the lethal
action of 1080 is due to its conversion in animal tissues to (-)
erythrofluorocitric acid. Once 1080 has been converted, it is

no longer susceptible to detection by current analytical methods.
It follows that an animal could die from a minimum/ lethal dose
of 1080 that would not leave detectable amounts of 1080, if all
of the ingested dose were converted to fluorocitrate before the
animal died.
In sum because the confidence limits for quantitative esti-
mates of 1080 residues in various animal tissues have not been
established and the reproducibility of such estimates has not
been determined, the use of chemical analysis results must be
interpreted with caution when assessing hazards.
Published toxicity data are also used frequently to assess
the hazards of 1080 to non-target species. Much of the tox-
icity data is based on studies by Ward and Spencer (1947)/ who
report toxicity figures for 44 species.
Examination of these data makes them suspect for use in
predicting potential risk to non-target wildlife. For example/
the magpie LD0 is reported to be 0.67 mg/kg and the LDioo to
be 1.3 mg/lag. According to the test report, the toxicant was
"fed" to a group of 12 magpies. No other details of the test
are reported. When a test substance is administered in Seed,
its toxicity is usually reported as a LC (lethal concentration)
rather them a LD (lethal do3e). More importantly, it is difficult
to determine how to use the data to predict hazards under field
conditions without details on the feeding schedule and the con-
centration of the toxicant in the feed. Another example from
this same paper is the reported lethal dose for the bobcat.
Five bobcats received an intraperitoneal injection of 1080
equivalent to .66 mg/kg. All animals died. While a dose of .66
mg/kg resulted in 100 percent mortality/ the data do not indicate
whether a lower dose would also produce similar results. More-
over/ it is difficult to use the data to predict hazards in the
field/ since the route of administration in the test differs
from that expected in the field/ where bobcats will eat 1080.
Most ot the other toxicity figures reported by Ward and Spencer
(1947) have weaknesses similar to those described in the above

These questions on toxicity and residue data are strength-
ened by examining studies on secondary poisoning by 1080 reported
by the U.S. Fish and Wildlife Service (Connolly, 1980). A dose
of 333 mg of 1080 (300 mg A.I.) was administered to a coyote.
After the coyote died, two groups of magpies were fed tissues
from this coyote. Neither group of birds was reported to exhibit
ill effects^ either during the experimental feeding or the obser-
vation period.
Sowever^ based on the assumed LD50 for the magpie of 1 mg/kg
(estimated by Connolly (1980) from Ward and Spencer's reported
LDq 0.67 rag/tog and LD^qq of 1.3 mg/kg) and the reported muscle
contents of 2.4 ppm 1080 from chemical analysis, at least some
mortality would have been predicted. At an LO5Q value of 1
mg/kg a 180 g magpie would have to ingest .13 mg of 1080 to
receive a median lethal dose. At the reported concentration in
muscle tissue and a daily consumption rate of 90 to 100 grams
per bird, each bird was consuming .216 to .24 mg of 1080 per"
day, slightly more than a median lethal dose per day. Five
birds ingested such amounts of poisoned coyote tissues daily for
7 days, but no mortalities resulted and no sublethal symptoms of
intoxication were seen.
Because of the uncertainties with toxicity data, the U.S.
Fish and Wildlife Service has initiated tests to determine the
toxicity of 1080 to magpies and other species. Xn his testimony
given at the information gathering hearings, Connolly stated,
"Trials in mid-July raised the possibility that magpies are more
resistant than we had thought." Although these tests are incom-
plete, and additional work is planned, test results to date
appear to underscore the questions raised on the reliabilty of
available toxicity values for use in evaluating hazard to
Questions have also been raised about the likelihood of
secondary poisoning from 1080. Dr. Sun of.the University of
California briefly addressed this point in his testimony at the
informational gathering hearings in Denver. He stated the

The question of secondary poisoning is essentially
no problem. Obviously if you have very large
amounts of fluoroacetate (1030) in the animal's
stomach# and some other animal eats that 3tomach»
happens to be eating fluoroacetate in large quan-
tities/ it is just the same problem as eating the
poison in the first place/ but in lethal doses of
fluoroacetate/ which is relatively small/ it is in
very/ very small quantities/ and fluoroacetate in-
gested is rather non-toxic. It undergoes a change
in the cytoplasm. I don't really understand the prob-
lem of secondary intoxication. If you eat very much
of something/ and another animal eats you, of course,
he gets intoxicated/ and it is not a real scientific
Dr. Kun in Appendix I of Connolly (1980) addressed this
point in somewhat more depth in the followingj
It should be recognized that (-) erythrofluorocitric
acid exhibits its unusual patent toxic action only
if it is biosynthesized in mitochondria by the
following reaction:
condensing enzyme
equ.l.: F-acetyl-CoA + Oxalacetate	^
(-) erythrofluorocitric acid + CoA.
If P-citrate were ingested, its toxicity is probably
negligible, because (-) erythrofluorocitric acid,
after entering cells from the blood stream, is
efficiently detoxified by the ubiquitous cytoplasmic
enzyme: ATP-cytrate lyase/ that cleaves F-citric.
The minute amounts of cytoplasmic F-acetate after
hydrolysis of F-acetyl CoA formed from ingested

is selectively consumed." Also, the report of cojotes dying one
to two weeks following 1080 application for rodent control (Hegdal
et al. 1979; Malloy, 1980) seems to raise questions with the
statement that "F-acetate in decaying tissues is likely to be
defluorinated in 5-10 hours to harmless glycolic acid + F- thus
a serious concern about 'epidemic* F-acetate poisoning through
poisoned carcasses appears unreasonable."
3. New Information - Bait Stations
In addition to the information discussed above (which per-
tains generally to the environmental hazards of 1080} EPA has
become aware of other information which pertains directly to the
hazards of 1080 use in large bait stations.
The uncertainties about the reliability of the analytical
methods used to detect 1080 residues in animal tissues raises
questions about the reliability of data presented in "The Cain
Report" showing residues in dead raptors. Although the analytical
method used is not specified, if it resulted in false positives,
the conclusions drawn from the data may over-estimate potential
problems. On the other hand, if the chemical method used gave
false negatives, conclusions drawn from the data could under-
estimate potential problems.
In testimony at the information gathering hearings in Denver,
Dick Randall, a representative of the Defenders of Wildlife,
provided new information on the potential hazards of 1080 bait
stations. Beginning in the fall of 1969 and continuing until
the predator toxicant ban in 1972, Randall monitored non-target
mortality associated with large 1080 bait stations. A tracer
material (zinc and cadmium sulfide) was combined with the 1080
poison used by the U.S. Fish and Wildlife Service in predator
control to identify animals killed by the baits. Of 32 animals
found dead and necropsied, 37 contained the tracer material
from the 1080 bait stations.

Table VI-3 shows the species which Randall found dead.
Neither Randall's written nor oral testimony specify the con-
centration in the baits*. However# in telephone contact with the
witness, he indicated that, although it was difficult, he had
made every effort to dose the bait at the recommended rate, 1.6g
1080 per 100 pounds o£ meat.
Lyle Crosby of the Wyoming Department of Agriculture also
gave testimony on a monitoring study conducted in Wyoming in the
winters of 1975-76 and- 1376-77 when over 1,000 1080 bait stations
were used. Baits were placed in every county of the state. In
oral testimony given in Denver, Mr Crosby indicated that one dog
was found dead in association with the control program; Mr.
Crosby also reported seeing two skunks which may have been killed
by 1080. In personal communications with the witness he indicated
that in total he was aware of one dog, two skunks and one badger
which were found dead in the vicinity of 1080 bait stations used
in Wyoming during the winters of 1975-76 and 1976-77. Ee also
indicated that some other species were found dead, but did not
know the specifics.
Finally, regarding the large bait stations, Suy Connolly of
the Denver Wildlife Research Center, in his testimony, suggested
that hazards could be reduced 'in the large bait station, while
still providing control, by reducing the 1080 concentration
from the previously recommended rate of 1.0 mg/oz. to 0.5 mg/oz.
of bait. According to Connolly, even at the reduced concentration,
4 ounces of bait would be lethal to most coyotes.
Examination of results of the feeding studies reported by
Robinson (1948) supports this suggestion. In order to determine
the effect of 1080 stations on non-targets, Robinson fed meat
from a bait station (treated at a rate of 1.6g 1080/100 lbs.
bait) to several non-target species. Table VI-4 shows the
results. These data were not discussed in "The Cain Report,"
though that report did reference other parts of Robinson's work.

Table 71-3* Species Pound Dead in Vicinity of 1080 Bait Station
Which Contained Tracer Materials.
Species	Numbers
Coyote	4
Dog	1
Badger	8
Bobcat	2
Pine martin	2
Mink	1
Weasel	1
Golden eagle	6
Red tailed hawk	1
Magpie	4
Prairie falcon	2
Unidentified hawk	2
Sharp-shined hawk	1
Canada jay	1
Rough-leg hawk	1
From 0. Randall's testimony, Denver.

Table VI-4. Results of Feeding Trials of Meat Treated at a Rate
o£ 1.6 ntg. of 1080 per 100 lbs. Bait.
Species	Number
Badger	2
Raccoon	1
Magpie	3
Magpie	7
Golden eagle	4
Ferruginous roughleg
hawk	2
Marsh hawk	2
American rough leg	1
Ferruginous roughleg
hawk	1
Prairie falcon	1
Marsh hawk	1
Amount Consumed
11 and 7 oz.
6 oz.
19.4, 11.0, &
10.9 grams
Each fed 4.S grams
10 # 131 16 & 22 oz.
All they could eat
All they could eat
Less than an oz.
Less than an* oz.
Less than an oz.
Less than an oz.
One badger died.
The other was left
with nervous
Left with nervous
All died
4 died
The one receiving
16 oz. died.
One died
One died
(Robinson 1948)

4. New Delivery Methods: Single Lethal Baits
Prior to the 1972 order, the use of single lethal baits con-
taining 1030 was not common. Since then, several individuals and
organizations have proposed this application method as an alter-
native to large 1080 bait stations to control canids (Beasom,
1976; Nesse, 1977; Wade, 1977; Wyoming Department of Agriculture/
1977; South Dakota Department of Agriculture, 1977; Colorado
Department of Agriculture, 1977; Ketron Inc., 1979; National Wool-
growers Association, 1981).
In general, the documents cited above propose three slightly
different use strategies for single lethal baits made of ground
or rendered lard, tallow, or other animal tissues. The following
is a summary of the three methods which have been proposed:
1.	Seasom (1976) proposed using baits weighing 9 grams
containing 2.9 mg 1080 each, systematically placed
on a 20 acre grid pattern, with baits located at
the most likely animal travel lane within 20 feet
of each grid intersection. This would equal 32
single lethal baits per square mile.
2.	Nesse (1977) proposed baits weighing 15 grams con-
taining 3.6 mg 1080 each to be placed around live-
stock or wildlife carcasses, coyote travel trails,
scent posts, den sites, or other "draw* stations.
A maximum of two baits at a single placement
location and use rates not to exceed 10 single
lethal baits per section was proposed, with bait
placement locations situated at least 1/4 mile
3.	Wade (1977.) recommended baits generally smaller
than 20 grams each containing S.O mg 1080 per
bait'iSfor coyotes, 3.0 mg 1080 ,per bait for red

foxes and 2.0 mg 1080 per bait for gray foxes.
Baits should be placed near established draw
stations or preferred travel routes in suitable
locations for the target animals to find them.
The number of baits placed at each station site/
normally from 10 to 30, is determined by field
inspection and by the history of target and non-
target species' activity around each station area.
It is useful to contrast the relative hazards of large bait
stations vs. single lethal drop baits, to the extent such
comparison can be made. Any comparison of the two delivery mech-
anisms is rendered difficult by uncertainty as to the density/
or number per unit area, of small baits used. Of the three
approaches described above the lowest density, suggested by
Neese, would permit 10 baits per section (one square mile). If
each bait contained 3.S mg# the total amount of 1080 used would
amount to 35 mg per section. The typical large bait station, in
contrast, would be a 60-80 lb. piece of horse or sheep containing
1.6 g of 1080 per 100 lb., or 1 mg of 1080 per ounce. The entire
station would contain 960-1280 mg of 1080. At a prescribed
density of 1 station per 36 sections, the 1080 used would equal
27-36 mg. per section, an amount comparable to that for single
lethal baits. The question becomes: Which poses more hazard,
one large bait in 36 square miles, or 360 small baits dispersed
over 36 square miles? (Connolly per. comm.).
Proponents of 1080 single lethal baits theorize that this
use would mitigate 1080's hazards to non-targets species. They
argue that the hazard to non-targets of primary poisoning is
reduced because each single lethal bait contains less than an
average lethal dose for many of the non-target animals at risk.
This, coupled with widely spaced placement of baits in treated
areas, reduces the chance of non-target species consuming a
lethal dose. Proponents of 1080 baits also contend that the
risk of secondary poisioning is reduced. Since this risk is
related to the amount, of 1080 consumed by the primary consumer,

and the amount of toxicant consumed by a primary consumer is
limited by the widely spaced placements and the amount of toxi-
cant in a bait, the single lethal dose presents a lower secondary
hazard than large meat bait stations. (Beasom, 1976; Xetron
Inc./ 1979)
While the arguments concerning the safety of drop baits are
logical, the hypothesis remains relatively untested. More impor-
tantly, the argument is based primarily on available toxicity
data which indicate that target species are more sensitive to
1080 than most non-target species. But, as pointed out earlier,
these toxicity data are of questionable reliability. Therefore,
any discussion of potential hazards of this use pattern must be
considered speculative.
Of particular interest beyond toxicity is selectivity of
acceptance of single lethal baits. Several factors appear to
influence this, including application rates, bait placement,
timing, bait acceptance by target and non-target species and
their relative densities. Although little research has been
initiated which specifically addresses the use of 1080 in single
lethal baits, some studies have been completed using other toxi-
cants in drop bait3. In addition, other studies using non-toxic
marking agents also provide insight into potential non-target
species exposure.
Beasom (1974), evaluating selectivity of various predator
control techniques in southern Texas, experimented with strych-
nine single lethal baits composed of chicken eggs, horsemeat and
pork fat. Approximately 2,000 eggs and 8,000 horsemeat or pork
baits were used over a two year period. The latter, which may
closely resemble the baits proposed for 1030, was placed on the
ground against or under grass at the side of study area roads.
Species found dead at or near the horsemeat or pork fat baits
were: 35 coyotes, 1 raccoon, 17 striped skunks, 1 spotted skunk,
5 opossum, 2 badgers, 2 cotton rats, 7 white-footed mice, 4
grasshopper mice, 1 pygmy mouse, 3 western harvest mice, 5 Harris'
hawks, 2 marsh hawks, 7 caracara, and 1 great homed owl. Spe-
cies reported dead near the strychnine egg baits were: 5 coyotes,

2 bobcats, 31 raccoons, 16 striped skunks, 4 opossum, 6 badgers,
2 collard peccaries, 6 armadillos, 36 cotton rats, 1 ground
squirrel, 17 white-footed mice, 15 grasshopper mice, 2 pygmy
mice, 7 western harvest mice, 3 marsh hawks, 13 caracara and 11
indigo snakes. The author also indicated other species could
have been killed and not found, especially in the case of raptors
which could pick up a bait and fly some distance away before
devouring it.
As part of a study evaluating an antifertility agent, infor-
mation was collected on consumption of drop baits by coyotes and
other species (Linhart et al., 1968). Rendered beef tallow
baits, aproximately 1/3 ounce with 1 percent seal oil rolled in
liver meal or blood meal, were placed on coyote sign along ranch
roads and at stock banks. Each station consisted of a smoothed
or sifted circle of sand and dirt in the center of which was
placed one to three test baits* Species consuming the bait were
determined by tracks left at the station. In six tests made in
New Mexico and Texas, 321 baits were eaten or carried off by
various animals. Of these, coyotes took 22 percent, rodents 52
percent, ravens 12 percent, and miscellaneous species 1 percent,
while the remaining 13 percent were taken by unidentified ani-
mals. Based on these tests the researchers believed that other
carnivores, with the possible exception of the skunk and fox,
seldom eat baits intended for coyotes, and that selective bait
placement, the relatively small number of baits per.square mile,
and the extended home range of coyotes also decrease the likeli-
hood of other carnivores eating baits.
Other tests conducted by the U.S. Pish and Wildlife Service
addressed the use of single lethal drop baits around draw sta-
tions (Tigner, 1981). Two tests were conducted in the winter of
1976-77, one near Rawlins, Wyoming and the other near Ft.
Sumner, New Mexico. Draw stations, consisting of half a sheep
carcass in Wyoming and about 1/4 of a cow carcass in New Mexico,
were located based on coyote sign with 20 small baits placed
around each station and replenished a week later. Twenty-four
draw stations were used in Wyoming and 19 in New Mexico. Each

small bait was made of lard/ coated with £ish meal/ and impregna-
ted with a physiological marker. Two weeks after the last bait-
ing, coyotes and other species were collected and examined for
presence of the markers. Following are the results of the tests
in Wyoming (animals sampled/ animals marked): deeraouse 8/8,
magpie 6/2, golden eagle 3/1, coyote 55/5, bald eagle 1/0 and
rough-legged hawk 1/0. In New Mexico the following were found:
swift fox 2/1, striped skunk 8/3, coyote 11/3, grasshopper mouse
5/1, greathomed owl 5/1, porcupine 6/0, hognosed skunk 1/0,
bobcat 1/0, red-tailed hawk 1/0, raven 2/0, vulture 3/0, deer
mouse 1/0, kangaroo rat 5/0, woodrat 1/0, marsh hawk 2/0, rough
legged hawk 1/0, spotted ground squirrel 1/0.
These studies suggest that several species would eat 3ingle
lethal baits intended for coyotes. However, in the absence of
reliable toxicity data potential impacts can not be further
Connolly, in hi3 testimony given in Denver, referenced recent
testa conducted by researchers at the Denver Wildlife Research
Center and Texas A a M University evaluating primary hazards of
single lethal baits to non-targets. Although details of the
tests were not presented, he indicated that both raccoons and
golden eagles had been fed single lethal bait containing about 3
mg of 1080. Although he indicated only a few animals had been
tested, most showed symptoms of Intoxication but all survived
and all appeared to have recovered fully. He went on to say much
additional work is needed to determine the hazard of single leth-
al dose coyote baits to all non-target species whose habits make
them likely candidates for exposure.
Also, the number of single lethal baits which non-target as
well as targets'would consume needs to be addressed. This would
appear to be related to the selectivity of baits, and baiting
strategies, neither of which is well defined. If the country-
side were saturated with single lethal baits, chance of exposure
to non-target species may be significant and may increase the
chance of multiple exposures as well. As the,number of baits at

a site is reduced! it seems reasonable that the chance of exposure
is reduced. However/ the use of draw stations may have the
tendency to concentrate some species. Under these conditions,
even at low application rates, exposure of non-target species
may result in individual mortality and affect populations, i.e.
wide-ranging carnivores and raptors, particularly eagles.
Secondary poisoning from 1080 when used in single lethal
baits is also of concern. As stated earlier, proponents argue
that secondary risk is related to the amount of 1080 ingested by
the primary consumer.
Again, although logical, several questions about this theory
need to be addressed before a complete assessment of potential
secondary hazard from single drop baits can be made. The ques-
tions include: How many baits would a target or non-target spe-
cies be likely to consume before death? Will the resulting
tissue 1080 levels be in the toxic range of scavengers? What
amount of 1080 will remain in the gastrointestinal tract? Once
again, answers to these questions depend on the existence of
reliable toxicity data and residue analysis methods.
5. New Delivery Methods: 1080 Toxic Collar
With the collar, exposure of either target or non-target
species to 1080 occurs only after a collar is punctured. After
a collar is punctured, the potential hazards associated with
other predacidal uses of 1080 are present, i.e., secondary poi-
soning of animals that scavenge remains of target coyotes, and
primary poisoning of animals that scavenge carcasses of coyote-
killed collared livestock.
Although most of the research to date has been on the effi-
cacy of the toxic collar, work by the.U.S. Fish and Wildlife
Service has addressed some of these potential hazards posed to
non-target species from the use o£ 1080 in the toxic collar
(Connolly, 1980).
Most of the research by the U.S. Fish and Wildlife Service
has addressed the potential for secondary poisoning of scavengers

which might feed on the remains of poisoned coyotes. Three
methods were used to investigate this ha2ard: 1) observations of
scavenging under field conditions; 2) analysis of 1080 residues
in poisoned carcasses; and 3) feeding poisoned coyote carcasses
to avian scavengers, specifically magpies. The latter two meth-
ods appear to show the most promise of providing information
which can be used to assess the potential for secondary hazards
from using 1080 toxic collars.
However, the usefulness of the data collected by these two
methods to assess hazard once again relies on the accuracy of
the chemical residue analysis and toxicity data for scavengers.
As discussed earlier, the data in these areas are not completely
reliable. Therefore# until the reliability of this data is
determined/ caution should be employed in drawing general conclu-
sions from the results of these tests.
With this in mind, the information reported by Connolly
(1980) is encouraging with respect to secondary hazard to avian
scavengers when 1080 is used in the toxic collar. The most
significant results come from studies in which magpies were fed
poisoned coyote carcasses or tissues.
Five groups of five magpies were confined with carcasses or
tissues from four coyotes blown to have been poisoned by 1080.
For the test/ magpies spent seven day3 of continuous confinement
with a carcass or tissues of a poisoned coyote. No other food
was available. This was followed by a seven day observation
period during which uncontaminated rations were fed to the mag-
pies. During the first trial/ researchers reported that magpies
fed heavily on the poisoned carcass/ but no birds showed any ill
effects. The coyote used in this test/ as in the second and
third/ had died after attacking a collared lamb in pen trials.
In the next two trials* which were run concurrently, four of
the 10 test birds died. Researchers concluded that the magpies
died of starvation* rather than 1080 poisoning. This conclusion
was based on two observations. First/ hot weather during the
trial dried out the coyote carcasses so- that the birds were p.
unable to feed for the entire seven day trial? and second/ the

average weight of the four dead birds, 107.4 g, was well below
that of four wild birds which were collected for comparison,
181.0 g. The researchers also noted that low 1080 residues were
found in breast muscle and gizzard from one test bird. They
suspected that this bird also died of starvation, as the fol-
lowing explains:
The bird that contained 1080 residues had been caged
with coyote *DM 385, in which the highest 1080 con-
centration was 0.27 ppm in the hip muscle. If the LD50
for magpies is 1 mg/log, this bird would have to eat
over 65Og of muscle form this coyote to have ingested
an LD50. Actually little feeding had occurred/ so
it is doubtful that the bird could have died of
secondary poisoning.
Que to the complications in the above test, the next trials
did not place an entire carcass with test birds. This trial
used coyote tissues which were dissected soon after death, with
muscle separated from all other soft tissues (heart# liver,
kidney, stomach, intestines, etc.). Tissues were refrigerated
until fed to birds. Daily feedings were adjusted to achieve
maximum voluntary consumption of about 90-100g per bird per day.
Also, in these tests the coyote was orally administered the
entire contents of one 30ml collar, 333mg of 1080. Two groups
of five magpies were fed tissues from this coyote. One group
was fed 500g of muscle tissue each day for seven days, the other
SOOg of soft tissue each day for two days. Researchers did not
observe ill effects in either group during the test or the sub-
sequent seven-day observation period.
These two trials appeared to subject magpies to substan-
tially greater risk of secondary poisoning than the magpies
might reasonably be expected to encounter under field conditions.
Again, questionable residue analysis and toxicity data hamper
interpretation of the results. If the reported residues in
tissues are accurate/ the dose received by magpies in these

tests was approximately two and a half times the highest residue
reported for a coyote which was known or suspected of being
killed by 1080 from the collar. If the magpie is as sensitive
to 1080 as available toxicity data indicate, secondary hazard to
other avian scavengers would appear limited from the use of 1080
in the toxic collar*
In addition to avian species, the potential for secondary
poisoning to mammalian species is of concern. No specific tests
are available which address this aspect of 1080*s use in the
toxic collar.
Possibly presenting greater non-target hazard are the car-
casses of sheep or goats with punctured collars. If both bladders
of the 30 ml collar are broken* 333 mg of the toxicant could be
present on and around the neck area of the dead sheep or goat
(twice this amount for the 60 ml collar). Even if available
toxicity data are in error by several times, this amount of 1080
may present a lethal dose for most scavengers. For example,
the turkey vulture, one of the least sensitive species to 1080
according to studies by Ward and Spencer (1947), could receive a
multiple median lethal dose from the neck area. The reported
LD50 of 1080 for the turkey vulture is 20 mg/log; hence, a median
letJial dose for a 1.6 kg bird is 32 mg, an amount which could
easily be present on the neck area of killed collared livestock.
Connolly (1980) presented the results of four trials which
were designed to assess this concern. One magpie test, using
the same approach as described for the assessment of secondary
hazard, was conducted. X coyote-killed lamb with a punctured
collar was caged with five magpies for seven days. The birds
scavenged heavily, but none were poisoned. Caged birds were
reported to have limited their feeding to lamb tissues exposed by
the killer coyote, and not on the neck and collar area. This
feeding pattern was indicated to be similar to what Connolly
observed for vultures in the wild.
In addition, three trials with domestic dogs were conducted.
They were allowed to scavenge at will on carcasses of coyote-
killed collared goats at field test sites in Texas. As soon as

a kill with, a punctured collar was found, a dog was taken to the
carcass and allowed to feed. The first two trials consisted of
one feeding^ while the third trial allowed the dog to feed once
or twice daily for nine days. In the three trials, no evidence of
poisoning was observed. Researchers concluded that apparently
little or no toxicant was consumed because the dogs did not feed
on or near ruptured collars.
Although these tests appear to indicate minimal hazard, the
tendency of test species not to feed near ruptured collars may
not be representative of feeding behavior of other scavengers.
Field observations from other studies indicate that in some
instances, particularly when food supplies are limited, soft
tissues of livestock carcasses are completely scavenged leaving
only hide and bones (Fite per3. com.). This suggests the possi-
bility of- scavengers ingesting the toxicant. Also, scavenging
of carcasses under field conditions is not limited to one spe-
cies; some species, such as the golden eagle, are capable of
penetrating the skin, exposing tissues. Such scavenging could
provide those, such as the magpie, which are not capable of
penetrating the skin, with other openings in which to feed,
possibly on contaminated portions.
A further indication of the potential for scavengers to be
poisoned by feeding on dead collared livestock is gained from
the cited U.S. Fish and Wildlife Service report on three coyotes
that died in pens where contaminated lamb tissues had been buried
(Connolly, 1980). Reseachers indicated that because of the
possibility that the coyotes could have been poisoned by 1080
contamination of the pens, the affected coyotes were analyzed
for the toxicant. One coyote proved positive for 1080. Re-
searchers speculated that this coyote had consumed contaminated
lamb remains cached by coyotes that had killed collared lambs.
In summary, the major potential hazards to non-target spe-
cies from the use of 1080 in the toxic collar are to avian and
mammalian scavengers from carcasses of poisoned coyotes and
carcasses of sheep or goats with punctured collars. Due to

scarity toxicity data and questions on the accuracy of quantita-
tive estimates of 1080 residue in tissues, the extent of these
potential problems cannot be fully assessed. Present results
are encouraging/ however.
3. Hazards to Humans
1. Information- Used In 1972 Decision:
Risks to man were a factor in the 1972 decision to cancel
the use of 1080 for predator control. Specifically, the Agency
19.	1080 is highly toxic to all species. The danger-
ous dose for man is 0.5-2 mg/kg. The chemical acts
rapidly upon the central nervous and cardiovascular
systems with cardiac effects. Effect is usually too
quick to permit treatment# and antidotes are relative-
ly valueless.
20.	According to one authority, prior to 1963 there
were 13 proven fatal cases, five suspected deaths,
and six nonfatal cases of 1080 poisoning in man,
although it is not clear to what extent predator
control materials were implicated.
During the period 1950 to 1972/ 1080 was being used both as
a predacide and as a rodenticide primarily to control field
mice/ ground squirrels/ prairie dogs, and other field rodents.
2. Summary of New Information
a. Use History Since 1972
Since 1972/ the only federally registered uses of 1080 have
been for the control of rats and mice. While the target pest

and the bait material (treated grain, oat groats, etc.) differ
from the predacide uses of 1080, the human exposure from preparing
the poisoned bait is roughly comparable. For the most part both
the 1080 rodenticide and predacide were prepared by state
personnel from concentrated 1080 purchased from the manufacturer*
The treated bait was also distributed and placed by state
personnel or people under their direct supervision.
Two aspects of the use of 1080 have changed. First, since
1978, 1080 has been classified as a restricted use pesticide.
(See 40 CPS S162.31.) As a result, 1080 may be used only by a
certified applicator—that is, a person who has been specifi-
cally trained in the safe use of highly toxic pesticides—or a
person under his direct supervision. (Prior to 1972 EPA lac iced
the authority to require that a pesticide be used only by certi-
fied applicators.) Second, the directions for using federally
registered uses of 1080 have been clarified and expanded so that
users applying the treated bait will know how to minimize risk
to humans and the environment.
EPA has reviewed the reports of pesticide poisonings submit-
ted since 1972 and has found seven incidents of human poisoning
involving 1080. None of these poisonings was fatal. In two
cases 1080 involvement is considered highly unlikely, and another
of the cases was an attempted suicide.
Guy Connolly testified at EPA's public hearings in Denver
that 1080 is used extensively in New Zealand (over 4,000 lbs.
per year) and that no human fatalities have occured from such
Another witness at the Denver hearings, Lyle Crosby, stated
that use of 1080 bait stations to control predators in Wyoming
in 1975-77 did not have adverse effects on any humans.
Another witness at the Denver hearings, V.M. Howard, testi-
fied that a 1080 filled collar broke when it was being removed
from a goat. Some of the 1080 contents splashed onto his hands
which he then washed. The collar was placed in a plastic bag
for disposal, with some of the 1080 spilling on the ground which
was subsequently covered with dirt. Howard said he suffered no

ill effects.
For the period 1966 to 1981/ SO incidents involving domestic
animals were reported. Over 110 dogs were killed, more than 10
dogs survived after treatment/ 30 other domestic animals (mostly
cats) were killed, and three other cats were affected but survived.
At least half of the dogs' deaths were the result of possi-
ble intentional poisonings (EPA, PIMS, 1981).
b.	Toxicity to Humans
Signs and symptoms of 1080 poisoning vary. However/ they
may be classified into three categories: CXS (central nervous
system)/ cardiac, and depression syndromes. The CNS syndrome is
characterised by hyperactivity, phonation, tonic spasms and
convulsions which lead to respiratory paralysis. The cardiac
syndrome is associated with blanching of the retina, muscular
weakness, clonic convulsions, and ventricular fibrillation.
The depression syndrome is associated with decreased activity,
respiratory depression and bradycardia. The time before the
onset of symptoms of poisoning almost always exceeds 1/4 hour
and death most often occurs between one hour and one day after
ingestion of a lethal dosage. The LD50 for humans has been
estimated to be approximately 2.0 mg/kg.
The development of pathologies is frequent, even after single
dosages. Definite histologic abnormalities in the myocardium
have been reported. Hemorrhagic changes in the liver/ heart,
aorta, and brain sometimes particularly occur in poisoned mammals.
Biochemically, 1030 is thought to exert its toxic effect by
inhibition or blocking of citrate and succinate metabolism
within the Kreb's cycle. The necessary biochemical transfor-
mations and the time required to impair functions account for
the lag between ingestion and the development of symptoms.
c.	Symptoms and Antidotes
1080 is absorbed rapidly from the gastrointestinal tract
and symptoms may not appear for 30 minutes or more. Therefore,

after an accidental poisoning, first aid measures should be
started as soon as possible. Vomiting must be induced without
delay to remove any unabsorbed 1030 in the stomach.
A physician can provide several beneficial treatments. In
addition to stomach lavage followed by purging, the physician
can institute acetate therapy/ which is to some degree antidotal.
Monacetin, sodium acetate, ethanol, and acetamide have been
recommended. Monacetin appears also to act directly to counter-
act cardiac arrythmias, which could lead to death in humans.
Monacetin would be the antidote of choice if available. Zt is
unfortunately generally unavailable and therefore has not been
used in the treatment of humans. Barbiturates are used to ame-
liorate convulsive episodes. Continuous cardiac monitoring is
A true antidote to 1080 poisoning is not available. If
treated immediately, however, and if the amount of poison in-
gested is not too large, first aid measures (induced vomiting,
stomach lavage, purging) are effective.

Albright, S. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxi-
cants. U.S. Environmental Protection Agency, July 28-29,
1981, Denver, Colorado, and July 31, 1981, Washington, O.C.
Althoff, D.P. and Gipson, P.S. (1981) Coyote family spatial re-
lationships with reference to poultry losses. Journal of
Wildlife Management, 45, 641-649.
Anderson, T.E., Caroline, M. and Beavers, J.L. (1974a) An evalua-
tion of aerial hunting as a means of protecting sheep &
goats from coyote & bobcat predation in Uvalde and Kinney
Counties, Texas. Unpublished report, U.S. Department of
the Interior, Fish and Wildlife Service.
Anderson, T.E., Ellard, H.D. and Caroline, M.S. (1974b) A prelim-
inary comparison of the effectiveness of traps, preda-
cides and aerial hunting in protecting calves from coyote
predation in Gray County, Texas. Unpublished report, U.S.
Department of the Interior, Fish and Wildlife Service.
Andrus, C.D. (1979) Memorandum on animal damage control to Assis-
tant Secretary of the Interior for Fish and Wildlife and
Paries. U.S. Department of the Interior.
Andrus, C. (1980) Keynote address processing of the predator
control summit. January 15, 1980, Austin, Texas.
Animal Damage Control Policy Study Public Hearings. (1973) U.S.
Department of the Interior. Boise, Idaho, and Washington,
D.C., May, 1978.
Aplin, T.E.S. (undated) Poison plants of Western Australia. West-
ern Australia Department of Agriculture, Bulletin 3772,
Arena, J.M. (1970) Poisoning-toxicology-symptoms-treatment. 2nd
ed. (Springfield, Illinois: Charles C. Thomas Publisher,
1970), 715pp.
Armentrout, D. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental^ Protec-
tion Agency, Denver,.Colorado, July 29, 1981, 340-362pp.
Ashorn, A. (1981) Material submitted in conjunction with informa-
tion gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.

ASTM Predator Task Group. (1976) New standard guideline for the
use and development of sodium monofluoroacetate (Compound
1080) as a predacide. Revision of 4/15/76.
Atkins, N. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Washington, D.C., July 31, 1981, 71-89.
Atzert, S.P. (1971) A review of sodium monofluoroacetate (Compound
1080) its properties, toxicology, and use in predator and
rodent control. a.S. Department of the Interior, Fish and
Wildlife Service Special Scientific Report--Wildlife Ho.
146. Washington, D.C., 34 pp.
Baker, J* (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 23-29 1981,
Denver, Colorado, and July 31, 1981, Washington, B.C.
Balser, D.S. (1974) A review of coyote control research. Pro-
ceedings; Sixth Vertebrate Pest Conference, Fresno Cali-
fornia, 171-177.	——— -
Barron, J. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Washington, D.C., July 31, 1981, 152-169.
Beasom, S.L. (1974) Selectivity of Predator Control Techniques
Techniques in South Texas. Journal of Wildlife Management
38(4): 837-844.	:
Beasom, S.L. (1976) Alternative Use Patterns of 1080 Bait for
Effective Predator Control While Reducing Environmental
Hazard. Texas A & M University.
Beck, J. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 23, 1981, 84-97.
Sekoff, M. (1975) Predation and aversive conditioning in coyotes.
Science 187: 1096 pp.
Boddicker, M.L. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Washington, D'.C., July 31, 1981, 194-212.
Bourne, J. and Dorrance, M.J. (1981) Lithium chloride aversion
for coyote predation of domestic sheep. Unpublished manu-
script .
Bourret, L. (1981) Testimony at information gathering hearings
on predator control toxicants, U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981, 274-238.

Bowns, J.E. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver/ Colorado, July 28, 1981, 104-115.
Brawley, K.C. (1977) Domestic sheep mortality during and after
tests o£ several predator control methods. M.S'. thesis,
University of Montana, Missoula, 66 pp.
Bums, R.J. (1977) Conditioned prey aversions and transfer of
avoidance to offspring in coyotes. Unpublished manu-
script. U.S. Department of the Interior, Fish and Wild-
life Service, Wildlife Research Center, Denver, Colorado,
13 pp.
Burns, R.J. and Connolly, G.E. (1980) Lithium chloride bait aver-
sion did not influence prey killing by coyotes. Proceed-
ings: Ninth Vertebrate Pest Conference, March 4-6, 1980.
Fresno, California, 200-203 .
Cain, S.A., Kadlec J.A., Allen D.L., Cooley R.A., Ho mocker M.G.,
- Leopold A.S. and Wagner F.H. (1972) Predator Control-1971,
Report to the Council on Environmental Quality and the
Department of Interior by the Advisory Committee on Preda-
tor Control. Inst. Environmental Quality, University
of Michigan.
Cargile, J.S. (4981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981.
Chenoweth, M.B. (1949) Mcnofluoroacetic acid and related com-
pounds. J. Parmacol. and Exptl. Therapeutic 97: 383-424 pp.
Colorado Department of Agriculture. (1977) Application for Regis-
tration of 1080 (EPA File Symbol 33968-T).
Colorado Department of Agriculture. (1981) Pending application
for registration of 1080 as predacide.
Connolly, G.E. Use of Compound 1080 in Livestock Keck Collars to
Kill Depredating Coyotes - A Report on Field and Labora-
tory Research. USDI, FWS, DWRC. (November 1978 - March
Connolly, G.E. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981.
Connolly, G.E., Griffiths, R.E. and Savarie, P.J. (1978) Toxic
collar for control of sheep-killing coyotes: A progress
report. Proceedings: Eighth Vertebrate Pest Conference,
Sacramento, California, 197-205.

Connolly, G.E. and Longhurst, W.M. (1975) The effects of con-
trol on coyote populations; A simulation model. Division
of Agricultural Science, University of California, Bulletin
1872, 37 ppl
Conover, M.R., Prancik, J.6. and Miller, O.E. <1977) An experi-
mental evaluation of aversive conditioning for control-
ling coyote predation. Journal of Wildlife Management
775-779.	'
Coppinger, L. (1980) So firm a friendship. Natural History, ,
Crosby, L. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 235-242.
Cruickshank, S. and Boyd, S. (1981) Material submitted in con-
junction with the information gathering hearings on predator
control toxicants. U.S. Environmental Protection Agency,
July 28-29, 1981, Denver, Colorado, and July 31, 1981,
Washington, D.C.
Data Resources, Inc. (1981) Death loss of cattle and calves
from all causes 1950 to 1980. Computer printout of
historical USDA statistics. Washington, D.C.
de Calesta, D. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
de Calesta, D.S. (1978) Documentation of livestock losses to
predators in Oregon. Oregon State University Extension
Service. Spec. Rep. in USDI-FW3 1978.
DeLorenzo, D.G., and Howard, Jr. V.W. (1976) Evaluation of sheep
losses on a range lambing operation without predator
control in southeastern New Mexico. Pinal Report. U.S.
Pish and Wildlife Service, Denver Wildlife Research Center
Contract 14-16-008-830.
Dorrance, M.J. (1980) Use of toxicants for coyote control by
livestock producers in Alberta. Proceedings? Ninth
Vertebrate Pest Conference, Fresno, California, 209-214.
Dungan, G. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 29, 1981, 462-468.
Early, J.O., Roetheli, J.C., and Brewer, G.R. (1974a). An
economic study of predation in the Idaho range sheep
industry, 1970-71 production cycle. Idaho Agr. Res.
Prog. Rep. No. 182. 49 pp.

Early^ J.O. (1974b) An economic study of predation in the
Idaho range sheep industry, 1972-73 production cycle,
Idaho Agriculture Research Project Report No. 186. 4 pp.
Ellins, S.R., Catalano, S.M. and Schechinger, S.A. (1977)
Conditioned taste aversion: A field application to
coyote predation on sheep. Behavioral Biology 20: 91-95.
Ellis, M.1C. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Faulkner, E.K. and Tigner, J.R. (1977) Birth rates of sheep
from range operations in Carbon County, Wyoming. Wyoming
Agricultural Extension Service B-643, University of Wyoming,
14 pp.
Garcia, J., Hankins, W.6. and Rusiniak, K.W. (1974) Behavioral
regulation of the milieu interne in man and rat. Science
185s 824-831.
Gee, C.K. and Magleby, R. (1976) Characteristics of sheep
production in the western United States. Agriculture
Economic Report 345, U.S. Department of Agriculture,
Economic Research Service, Washington, D.C. 47 pp.
Gee, C.K., Nielsen, D.B. and Stevens, D.M. (1977b) Factors in
the decline of the western sheep industry. Agriculture
Economic Report 377, U.S. Department of Agriculture,
Economic Research Service, Washington, D.C. 31 pp.
Gee, C.X., Magleby, R.S., Bailey, W.R., Gum, R.L. and Arthur,
L.M. (1977) Sheep and lamb losses to predators and other
causes,in the western United States. Agricultural
Economic Report Wo. 369, U.S. Department of Agriculture,
Economic Research Service, Washington, D.C., 41 pp.
Gee, C.K. (1979) Cattle and Calf Losses to Predators - Feeder
Cattle Enterprises in the United States. Journal of
Range Management. Vol 32, No. 2, March 1979, p 152-154.
Gilbert, R. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Glosser, J. (1981) Testimony at information gahtering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 47-61 and 247-253.
Green, J.S., and Woodruff, R.A. (1980) Is predator control
going to the dogs? Rangelands 2:187-189.

Green, J.S., Tueller, T.T. and Woodruff, R.A. (1980) Predator
control, economics and guarding doers. Ranaelands 2:
Grieb, J. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 21-28.
Griffiths, H.2. (1978) Problems encountered in averting captive
coyotes to sheep with lithium chloride. Unpublished
manuscript. U.S. Department of the Interior, Fish and
Wildlife Service, Wildlife Research Center, Denver,
Colorado, 6 pp.
Griffiths, R.E., Connolly, G.E., Burns, R.J. and Sterner, R.T.
(1978) Coyotes, sheep and lithium chloride. Proceedings;
Eighth Vertebrate Pest Conference, 190-196.
Grobel, L.O. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Gustavson, C.R., Garcia, J., Hankins, W.G. and Rusiniak, K.W.
(1976) Coyote predation control by conditioning. Science,
Gustavson, C.R., Kelly, D.J., Sweeney, M. and Garcia, J. (1976)
Prey-lithium aversions I: Coyotes and wolves. Behavioral
Biology 17: 61-72.
Gustavson, C.R., Sweeney, M»J., Brewster, R.G., Jowsey, J.R.,
and Hiligan, D.N. (1977) Taste aversion control of
coyote predation in Washington, California and
Saskatchewan. Unpublished manuscript, 11 pp.
Guthery, F.S. and Beasom, S.L. (1973) Effectiveness and
selectivity of neck snares in predator control. Journal
of Wildlife Management 42s 457-459.
Sayden, W. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver Colorado, and July 31, 1981, Washington, D.C.
Havens, M. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 29, 1981, 403-407.
Hegdal, P., Gutz, T., Fagerstone, K., Glahn, J. and Matsuhke, G.
(1979) Hazards to Wildlife associated with 1080 Baiting
for Ground Squirrels. U.S. Department of the Interior,
Fish and Wildlife Research Center, Denver, Colorado.

Hegdal, P., T. Gutz, and E. Fite (1980) Secondary Effects of
Rodenticides on Mammalian Predators. Proceedings:
Worldwide Furbearer Conference, Frostburg, Maryland.
Eelle, J. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency» Washington, D.C., July 31* 1981, 138-182*
Henderson# F.R. (1972) The extension trapper system in Kansas.
Proceedings: Fifth Vertebrate Pest Conference, Fresno*
California, 171-177. ;
Henne, F.R. (1975) Domestic sheep mortality on a western Montana
ranch. M.S. thesis, University of Montana, Missoula.
S3 pp.
Henne, F.R. (1977) Domestic sheep mortality on western Montana
ranch, p. 133-146. In: Phillips R.L. and Jonkel C. eds.
Proc. 1975 Pred. Symp. Montana Forest and Conservation
Experiment Station, School Forest., University of Montana,
Sibbard, C.T. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981, 148-158.
Sines, J. Material submitted in conjunction with the information
gathering hearings on predator control toxicants. U.S.
Environmental Protection Agency, July 28-29, 1981, Denver,
Colorado, and July 31, 1981, Washington, D.C.
Hodder, E. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 29, 1981, 340-362.
Hoff, C. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Washington, D.C., July 31, 1981, 46-60.
Hotchkiss, J. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Washington, D.C.
Howard, B.C. (1981a) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981, 254-261.
Howard, E.C. (1981b) Statement concerning compound 1080, a
predacide. Submitted to U.S. Environmental Protection
Agency for record for preliminary hearings on predator
control toxicants. Texas Sheep and Goat Raisers
Association, Texas Animal Damage Control Association,
Inc. Corrected copy, August 4, 1981.

Janklow, W. (1981) Material submitted in conjunction with the information
gathering hearings on predator control toxicants. U.S.
Environmental Protection Agency, July 28-29/ 1981, Denver,
Colorado, and July 31, 1981, Washington, O.C.
Jensen, R.J., Tobiska, I.W., and Ward, J.C. (1943) Sodium
fluoroacetate (Compound 1080) poisoning in sheep.
Amer. J. Vet. Res. 9: 370-372 (Cited by A&zert 1971).
Kaye, S. Handbook of emergency toxicology* (1970) 3rd ed.
(Springfield, Illinois: Charles C. Thomas Publisher), .
514 pp. (Cited by Atzert, 1971).
Ketron, Inc. (1979) Assessment of the Environmental Effects of
Predator and Rodent Control Programs in Wyoming Using
Strychnine and 1080. USDA (APSIS) Contract
Kirkbride, K. (1981) Material Submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Klataske, R. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 29, 1981, 377-389.
Klebenow, D.A. and McAdoa, K. (1976) Predation on domestic
sheep in northwestern Nevada, J. Range Manaqe. 29(2):
Knowlton, F.F. (1972) Preliminary interpretations of coyote
population mechanics with some management implications.
Journal of Wildlife Management 36; 369-382.
Koch, K. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 269-274.
Kosesan, W.S. (1981) Material submitted in conjunction with
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Sun, E. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Snvironmenttal Protection
Agency, Denver, Colorado, July 28, 1981, 97-104.
Larson, G.E., Wallace, M.3., Lewis, ,J.M. and Mansfield, M.S.
(1975) Coyote predation in sheep. In Update 75. A
Research Report of the Dixon Springs Agriculture Center.
University of Illinois, Urbana, DSAC, August 3.

Lazarus, M. (1956) The toxicity and relative acceptability of
some poisons to the wild rabbit Qryctolagus cunicuius.
CSIKO Wildl. Res. 1(2): 96-100, (Cited by Atzert, 1971).
Lehner, P.N. and Horn, S.W. (1977) Effectiveness of physiological
aversiva agents in suppressing predation on rabbits and
domestic sheep by coyotes. Pinal research report to U.S.
Pish and Wildlife Service, Colorado State University, Fort
Collins, Colorado, 104 pp.
Lehner, P.N., Krumm, R. and Cringan, A.T. (1976) Tests for
olfactory repellents for coyotes and dogs. Journal
of Wildlife Management 40: 145-130.
Leopold, A.S., Cain, S.A., Cot tain, C.M., Gwrielson, I.N., and
Kimball, T.L. (1964) Predator and Rodent Control in the
United States. Trans N. Am. Wildl. and Mat. Res.
Conf. 29.
Levinston P. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981, 213-221.
Linhart, S.B., Brusman, H.H., and Balser, D.S. (1968) Field
Evaluation of an Antifer-tility Agent, Stilbestrol, for
Inhabiting Coyote Reproduction. Trans. N. Am. Wildl and
Nat. Res. Conf. 33: 316-327.
Linhart, S.3. and Robinson, W.B. (1972) Some relative carnivore
densities in areas under sustainded coyote. Journal of
Mammalogy 53: 880-884.
Linhart, S.B., Sterner, R.T., Carrigan, T.C. and Henne, D.R.
(1979) Romondor guard dogs reduce sheep losses to coyotes:
A preliminary evaluation. Journal of Range Management 32:
Lynch, 6.W. and Nass, R.D. (1981) Sodium monofluoroacetate
(1080): Relation of its use to predation on livestock
in western National forests. Journal of Range
Management 34: 421-423.
Madsen, J. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981, 199-208 pp.
Malloy, J. (1980) Effects of 1080 Control of Columbian Ground
Squirrels on Target and Non-target Mammals and Bird
Populations. Summary report submitted to EPA by the State
of Montana.

Marcoux, R. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental
Protection Agency, Denver, Colorado, July 28, 1981.
McOmber, G. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 28, 1981.
Meduna, R.E. (1977) Relationships between sheep management
and coyote predation. M.S. thesis, Kansas State Univ.,
Manhattan, 1'40 pp.
Meike, D. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 65-78.
Montana Department o£ Livestock. (1981) Pending application
for registration of 1080 in single lethal dose baits.
Morris, J. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 308-313.
Mudd, D. (1981) Material submitted in conjunction with the information
gathering hearings on predator control toxicants. U.S.
Environmental Protection Agency, July 28-29, 1981, Denver,
Colorado, and July 31, 1981, Washington, D.C.
Munoa, J.a. (1976) Causes of sheep mortality at the Cook Ranch,
Florence, Montana, 1975-76. M.S. thesis. University of
Montana, Missoula. 55 pp.
Mass, R.D. (1977) Mortality associated with sheep operations in
Idaho. J. Range Manage. 30(4): 253-258.
Nass, R.D. (1980a) Livestock husbandry practices and their impact
upon predation: a literature review. U.S.. Department of
the Interior, Fish and Wildlife Service> Denver Wildlife
Research Center, 27 pp.
Nass, R.D. (1980b) "A list of husbandry practices that may reduce
predation" and "Husbandry practice trade-offs." Unpublished
tables, U.S. Department of the Interior, Fish and Wildlife
Service, 3 pp.
Nass, R.D. (1980c) Efficacy of predator control programs. Pro-
ceedings: Ninth Vertebrate Pest Conference, Fresno,
California, 205-208.
National Wcolgrowers Association, Inc. (1981) Letter dated
February 25, 1981 to Ann McGill Gorsuch, Administrator,
Environmental Protection Agency, Washington, D.C.

Natural Resources Defense Council/ Inc., Defenders of Wildlife,
Friends of the Earth, The Humane Society of the United
States, National Audubon Society, Inc., Hew York Zoological
Society, and Sierra Club. (1971) Petition requesting the
suspension and cancellation of registration of sodium
monofluoroacetate (1080), thallium sulphate, strychnine
and cyanide. Petition to U.S. Environmental Protection
Agency, 56 pp plus appendices and attachments.
Nesse, G.E. (1973) Predation and the sheep industry in Glenn
County, California. M.S. thesis, Univ. Calif., Davis
127 p.
Nesse, G.E., W.M. Longhurst, W.M., and Howard, W.E. (1976)
Predation and the sheep industry in California, 1972-
1974. Univ. Calif. Div. Agric. Sci. Bull. 1878. 63 pp.
Nesse, G.B. (1977) The Proposed Use of Single Lethal Sodium.
Monofluoroacetate (1080) Baits as means of Reducing
Livestock, and Wildlife Losses of Coyotes. Montana
Departments of Livestock, Pi3h and Game Agriculture, and
Health and Environmental Sciences; U.S. Fish and Wildlife
O'Brian, J. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Parker, R. (1981) Material submitted in conjunction with the information
gathering hearings on predator control toxicants. U.S.
Environmental Protection Agency, July 28-29, 1981, Denver,
Colorado, and July 31, 1981, Washington, D.C.
Pattison, F.L.M. (1959) Toxic aliphatic fluorine compounds.
Amsterdam: Elsevier Publishing Company, 227 pp.
Popoulas, J. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 29, 1981, 526-530.
Randall, D. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 29, 1981, 446-456.
Reed, N. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Washington, D.C., July 31, 1981, 7-27.
Reynolds, R.N. and Gustad, O.C. 1971. Analysis of statistical
data on sheep losses caused by predation in four western
states during 1966-69. U.S. Bur. Sport Fisheries Wildl.
Mimeo. 20 pp.

Robel, R.J. (1981) Testimony for information gathering hearings
on predator control toxicants. Submitted to U.S.
Environmental Protection Agency, July 10, 1981 plus
attachments, 5 pp.
Robel, R.J., Dayton, A.D., Henderson, F.R., Meduna, R.L., and
Spaeth, C.Mr. (1981) Relationships between husbandry
methods and sheep losses to canine predators. Journal of
Wildlife Management 45: in press.
Robinson, W.B. (1948) Thallium and compound 1080 impregnated
stations in coyote control. Journal of Wildlife
Management 12: 279—29S.
Robinson, W.B. (1953) Coyote control with Compound 1080 stations
in national forests. J. Forestry 51(12): 880-885, (Cited
by Atzert, 1971).
Robinson, W.H. (1970) Acute toxicity of sodium monofluoroacetate
to cattle. Journal of Wildlife Management 34(3): 647-648,
(Cited by Atzert, 1971).	'
Rogers, J.G. (1978) Repellents to protect seed crops from
vertebrate pests: Some considerations for their use
and development.- In: Bullard, R.W. (ed.) Flavor
chemistry of animal foods. American Chemical
Society, Symposium Series, Washington, O.C.,150-165.
Rost, G. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 261-269.
Ruckelshaus, William D. (1972) PR Notice 72-2, Re: Suspension
of Registration^ of Certain Products Containing Sodium
Fluoroacetate (1080), Strychnine and Sodium Cyanide.
Ryden, H. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency. Denver, Colorado, July 29, 1981, 499-512.
Schwartz, C. (1981) Material submitted in conjunction with the
information gathering hearings on predator control toxicants.
U.S. Environmental Protection Agency, July 28-29, 1981,
Denver, Colorado, and July 31, 1981, Washington, D.C.
Scott, C. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 29, 1981, 440-446 pp.
Shelton, M. (1972) Predator losses in one flock of sheep and
goats. Natl. Woolgrower 62:20.

South Datota Department of Agriculture. (1977) Application for
Registration of 1080 (EPA File Symbol 13808-U).
Stephens, W. P. (1981) New Mexico Department of Agriculture
Submitted for the record during public comment period on
compound 1080 predator hearings. August 4, Washington,
Sterner, R.T. and Shumake, S.A. (1978) Bait-induced aversions
in predators: Some methodological issues. Behavioral
Biology 22: 565-566.
Stevens, C. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Washington, D.C., July 31, 1981, 60-70 pp.
Stream, L. (1976a) 1976 lithium chloride taste aversion experiment
in Whitman County, Washington. Pinal report to Washington
State Department of Game, Olympia, Washington. (1976),
18 pp.
Stream, L. (1976b) Amendments and reconsiderations of the 1976
lithium chloride taste aversions program in Whitman
County, Washington. Submitted to Washington State
Department of Game, Olympia, Washington, (1976b), 7 pp.
Strojny, S. U.981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protec-
tion Agency, Denver, Colorado, July 29, 1981, 419-427.
Strom, R. E. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981, 313-318.
Taylor, R.G., Workman, J.P., and B'owns, J.E. (1978) The economics
of sheep predation in southwestern Utah. M.S. thesis.
Terrill, C.E. (1980) Trends of Predator Losses of Sheep and
Lambs from USDA Mortality Statistics. ASTM. March 7, 1980,
Fresno, California.
Tigner, J.R. and Larson, G.E. (1977) Sheep losses on selected
ranches in southern Wyoming. J. Range Manage. 30(4):
Tigner, J.R., Lawson, G.E., Roberts, J.D. and Johns B.S. (1981)
Progress Reports, Development and Evaluation of Baits and
Baiting Techniques for Field Application of Predacides.
U.S. Fish and Wildlife Service, Denver Wildlife Research
Center, Denver, CO.

Tirnrn, R.M. and Connolly, G.E. (1977) 3ow coyotes kill sheep.
Rangeman'3 Journal 4s 106-107.
Treat/ A. E. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency^ Denver/ Colorado/ July 28/ 1981/ 290-293.
Tucker, R.K., and O.S. Crabtree. (1970) Handbook of toxicity
of pesticides to wildlife. Bureau of Sport
Fisheries and Wildlife/ Denver Wildlife Research
Center/ Resource Publication No. 84/ 131 pp.
Turner/ R. (1981} Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental Protec-
tion Agency/ Denver# Colorado, July 29, 1981/ 436-440 pp.
Uhalde* G. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental
Protection Agency, Denver/ Colorado, July, 29, 1981/
U.S. Department of Agriculture. ESCS-AMS. (1978) Livestock and
meat statistics. Statistical Bulletin No. 522.
Washington, D.C.
U.S. Department of the Interior. (1978) Predator damage in
the west: A study of coyote management alternatives.
Fish and Wildlife Service publication, 168 pp.
U.S. Department of the Interior. (1979a) Mammalian predator
damage management for livestock protection in the
western United States. Final Environmental Impact
Statement, U.S. Fish and Wildlife Service Animal
Damage Control Program, 208 pp, plus comment*..
U.S. Department of the Interior. (1979b) The toxic collar for
selective removal of coyotes that attack sheep. U.S. Fish
and Wildlife Service, Denver Wildlife Research Center,
4th, 5th, and 6th progress reports under EPA Experimental
Use Permit No. 6704-EDP-14, 47 pp.
Vandehey, T. (1981) Material submitted in conjunction with the information
gathering hearings on predator control toxicants. U.S.
Environmental Protection Agency, July 28-29, 1981, Denver,
Colorado, and July 31, 1981, Washington, D.C.
Wade, D.A. (1977) Standard Guidelines for the Use and Develop-
ment of Sodium Monofluoroacetate (Compound 1080) as a
Predacide (ASTM Designation E590-76). Test Methods for
Vertebrate Pest Control and Management Materials, ASTM STP
625, W.B. Jackson and R.E. Marsh, Eds. American Society
for Testing and Materials, 157-170.

Wade, D.A. (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental
Protection Agency# Washington, D.C., July 31, 1981,
Wagner, F.A. (1972) Coyotes and sheep: Some thoughts on ecology,
economics, and ethics. 44th Utah State University
Faculty Honor Lecture, Logan, Utah, 59 pp.
Wagner, F.H. (1975) The predator control scene as o£ 1974.
Journal of Range Management 28: 4-10.
Walther, W.H., Williamson, P.M., Humphrey, M.D. and Johnson,
¦ S.L. (1979) 1979 Texas sheep and goat death losses and
marketing practices. Texas Crop and Livestock Reporting
Service, Texas Department of Agriculture, 12 pp.
Ward, J.C., and Spencer D.A. (1947) Notes on the pharmacology
of sodium fluoroacetate-Compound 1030. J. Amer.
Pharmaceutical Assoc. 36(12): 59-62.
Wentz, A. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency. Washington, D.C., July 31, 1981, 27-46.
Wilson, 7. (1981) Testimony at information gathering hearings on
predator control toxicants. U.S. Environmental Protection
Agency, Denver, Colorado, July 28, 1981.
Wintch, Mary (1981) Testimony at information gathering hearings
on predator control toxicants. U.S. Environmental
Protection Agency, Washington, D.C., July 31, 1981.
Wyoming Department of Agriculture. (1977) Application for.
Ragistration of 1080 (EPA File Symbol 35978-E)'
Wyoming Department of Agriculture. (1981) Pending application
for registration of 1080 as a predacide.
Young, S.P. and Dobyns, H.W. (194S) Coyote control by means
of den hunting. U.S. Department of the Interior,
Fish and Wildlife Service, Circular 7. Washington,
D.C., U.S. Government Printing Office, 8 pp.