Below is the uncorrected machine-read text of this chapter, intended to provide our own search engines and external engines with highly rich, chapter-representative searchable text of each book. Because it is UNCORRECTED material, please consider the following text as a useful but insufficient proxy for the authoritative book pages.
Research Concerns Numerous problems confront researchers who study feral horses, and two major logistical problems are inherent in the western United States. First, the rugged, mountainous nature and extensive size of most potential study areas create dilemmas in sampling study animals. It can be difficult to coordinate a large team of biologists to study each of the many popula- tions, although detailed life history and reproductive data can be gathered in this way (Berger, 1986). Depending on available funding, the only feasible method for gathering data over multiple, large study areas is aerial observation. Second, and more important, is the selection of suitable study areas. As pointed out by Berger (1986), several researchers were forced to terminate their projects because study animals were removed by the Bureau of Land Management (BLM) before the studies were completed. The BLM's mandate to manage feral horse herds has made difficult the selection of study areas that fulfill requisites for data collection. In Nevada this problem has been serious, and it has been necessary to select study areas that minimize these conflicts. Concerns about the conduct of field research activities have been addressed in the following five areas: â¢ Loss of the Clan Alpine horses; â¢ Problems and injury with the marker and radio collars; â¢ Foal orphaning and loss during roundups and aerial surveys; â¢ Abortion; and â¢ Disappearance of the penned animals at the Lovelock corrals. To an extent, these concerns reflect differing views of the ethical accept- ability of the methods used to study population control in wild and free- roaming horses. Examination of these ethical concerns, however, are be- 23
24 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY yond the professional expertise and charge of the committee. The commit- tee can respond to these issues only as they affect the quality of the research results. LOSS OF THE CLAN ALPINE HORSES In late August and early September 1987, BLM representatives and the research team began to round up animals from the Clan Alpine area for experiments. However, by September 1 the number of animals rounded up was insufficient for a satisfactory sample size. A decision was made to augment the Clan Alpine numbers with animals from the adjacent Augusta Mountain area. A fence separates the Clan Alpine and Augusta Mountain areas. On September 2 and 3, 133 animals were rounded up in the Augusta Mountain area and driven to a trap on the Clan Alpine side of the fence. Some horses were driven as far as 15 or 20 miles. The weather was hot and dry, and the horses were in relatively poor physical condition. Of the 133 captured animals, 42 mares were implanted with placebo capsules and equipped with radio collars, bringing the total number of pla- cebo-implanted mares in the Clan Alpine study area to 49. Another 33 were fitted with marker collars, bringing the total untreated, marker-collared popu- lation to 109. The animals were then released on September 2 and 3 in the vicinity of the trap. Between September 17 and October 30, 48 horses were found dead on the Clan Alpine side of the boundary fence. These included 17 radio- collared animals, 11 animals with marker collars, and 20 unmarked animals. A team composed of a veterinarian and BLM law enforcement and management personnel investigated the situation and concluded the following (U.S. Department of the Interior, 1987): The animals were attempting to return North to their home range and were prevented from doing so by the fence. Therefore, not knowing where water sources were located south of the fence, the animals walked the fence in both directions until they died from dehydration. The committee has received four other versions of the incident, including one account that there were no gaps in the fence and the horses were driven excessive distances around it. The committee deeply regrets this tragic incident, but it is in no position to impute culpability. It bears, however, the responsibility to assess the effects of this incident on the validity of the research results. The losses included 17 comparison mares with placebo implants. Another placebo-treated mare died about 12 hours after release, reducing the number of Clan Alpine comparison mares from the original 49 to 31. This reduced
RESEARCH CONCERNS 25 the statistical power of the test for the difference in that study area. Never- theless, the foaling rate of the 31 observed placebo implants in this area in 1989 (Table 2-5) was statistically higher than that of the treated mares. Moreover, for the study as a whole, there remained 30 comparison mares in the Wassuks in 1989. Thus the loss of the 18 Clan Alpine animals reduced the total number of comparison mares from 90 to 61. Given 2 years of observations on these animals and on the 1988 unmarked mares (Table 2-4), the loss did not render the research invalid or statistically inadequate. As discussed above, the foaling rates of the placebo implants in the different areas and years are consistently similar. Those rates are all statistically higher than the rates of the steroid-treated animals in their sec- ond and third years after implantation, and they are in a consistent, low range. COLLAR PROBLEMS The use of both radio and marker collars is a widely accepted practice in large-animal field studies. Based on extensive experience with these devices, no adverse effects were anticipated. Between 1987 and 1989, however, a number of horses involved in the study suffered injuries to their necks and ears that were caused by the collars used to locate and identify the experimental animals. Serious questions have been raised concerning the deaths of some of these animals, the nature and extent of the wounds, and possible changes in behavior of the animals as a result of collar problems. Other questions have focused on the design of the collars and the experience of the research team. Collars were attached to a total of 876 horses. Of these 876 animals, 336 received radio collars (291 mares and 45 stallions); the other 540 adult horses were fitted with marker collars. The radio-collared animals were part of the two fertility control studies. The marker-collared animals were intended to be part of a study of band stability; this study was discontinued due to a lack of funds. The collars will be removed by the BLM at the end of the fertility control study. Collar Design The collars are made from an industrial belting material composed of rubber and canvas that is inelastic but somewhat flexible. They are 4 inches wide, with a narrower section underneath the neck where the radios are attached. The collars are adjustable in circumference by means of a system of holes and studs, but fine adjustments are difficult or impossible. Attachment of the radio units stiffens the collars, making them quite inflexible over a span of several inches. All of the collars have 3-inch high numerals to
26 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY permit identification from the air. These collars are similar, although not identical, to ones employed successfully by members of the research team in an earlier wild horse census study. The belting material used for the collars in the present study appears to be less flexible than the material used previously. Collar Wounds Collars were first fitted in the fall of 1986, and problems arose in the spring of 1987. As soon as the problems became evident, the research team began searching for injured horses by helicopter with Dr. Gerald Peck, the on-site consulting veterinarian for the Nevada state office of the BLM. Animals that appeared from the air to have an injury needing treatment were darted and examined. The results are summarized in Table 4-1. During the years from 1987 to 1989, the research team darted a total of 77 animals. Collars were removed from 54 animals, which constituted 6 percent of the 876 collared animals. Twenty-seven were removed because they were too tight; 27 others were removed because they became too loose and slipped forward over the animals' ears. Six animals were subsequently recollared. A total of 48 animals had collars permanently removed. Twenty were radio-collared animals that formed part of the experimental population. Three additional radio collars slipped off animals in the field. In all, 23 animals were lost from the experimental population due to collar removal or loss. The other 28 collars that were removed were marker collars with no radio telemetry devices attached. Two other animals died while being captured to have collars adjusted or removed. The wounds caused by tight collars were unquestionably grim in appearance. In some cases, the horse grew into the collar material, so that the collar became imbedded in the animal's neck. In other cases, the collar abraded the skin under the neck where the radio unit was attached, causing an open sore that subsequently became infected. Loose collars rode up on the animals' necks and over their foreheads, causing sores on the ears. Primarily the injuries were surface wounds, which responded to treatment and healed. Dr. Peck told the committee that there was no sign of systemic infection, dehydration, or deterioration of body condition in the injured horses. Dr. Peck's diagnoses were communicated to BLM on January 25, 1988, in a report (see Appendix). Six of those animals (designated as recollared in Table 4-1) have subsequently had collars reattached. Dr. Peck, however, questioned the recollaring of horses with neck scar tissue; no animals were recollared after September 1988. The great majority of the neck wounds were associated with radio collars, rather than marker collars. Most of the wounds occurred under the neck, at
RESEARCH CONCERNS 27 TABLE 4-1 Darting and Collar Summary, 1987 to 1989 Number of Collared Horses Time Period and Study Area Horses Darted Collars Removed Collars Adjusted Collars Not Adjusted Horses Died Horses Recollared* June 1987- September 1988 Flanigan 11/0Â° 9/0 2/0 0/0 0/0 0/0 Beaty Butte 1/0 I/O 0/0 0/0 0/0 0/0 Wassuks 4/0 4/0 0/0 0/0 0/0 0/0 Stone Cabin 4/31 2/18 2/10 0/2 0/1 0/6 Clan Alpine 4/4 0/2 3/1 0/1 1/0 0/0 Subtotal 24/35 16/20 7/11 0/3 1/1 0/6 October 1988- October 1989 Flanigan 2/0 2/0 0/0 0/0 0/0 0/0 Wassuks 3/0 3/0 0/0 0/0 0/0 0/0 Stone Cabin 3/6 3/6 o/o 0/0 0/0 0/0 Clan Alpine 3/1 3/1 0/0 0/0 0/0 0/0 Subtotal 11/7 11/7 0/0 0/0 0/0 0/0 Total 35/42 27/27 7/11 0/3 1/1 0/6 "Number on left represents animals with collars over their ears; number on the right represents animals with tight collars. 6Six animals that had their collars removed were subsequently darted a second time so they could be recollared. the point where the radio unit was attached. In two cases where the collar had rotated, wounds were observed on the mane under the radio unit. Most observers attributed this effect to the stiffening of the collars at the point where the radios were attached. Dr. Peck also pointed out that the radio collars were put on with the horses lying down, tied up, and anesthetized, while the marker collars were put on with the horses wide awake, and standing in the chute with their neck muscles tensed. Stone Cabin Most of the problems with tight collars occurred at Stone Cabin. Collars were placed on 215 animals (101 of which were radio collared) in Septem- ber 1986. Between June 1987 and September 1988, 31 of those horses were subsequently identified as having collars that were too tight. The research team attributed the problems with tight collars at Stone Cabin to rapid weight gains. They stated that forage conditions were favorable, contraception was effective, and the animals experienced significant increases in body fat. BLM personnel generally agreed, pointing out that the horses at Stone Cabin
28 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY were collared during the second year of a 2-year drought, which was fol- lowed by 2 years of high precipitation and abundant forage production. Both the researchers and BLM personnel stated that the problem of collars becoming imbedded had not happened before and was not anticipated on the basis of previous experience. Age may also have been a significant factor. According to BLM personnel, many of the animals collared at Stone Cabin were young (2 to 5 years) and may have outgrown their collars in the normal process of maturation. One BLM official told the committee that virtually all of the horses from 2 to 5 years of age experienced collar problems. It is not possible to evaluate the effect of age in the absence of age distribution data. Dr. Peck also suggested that the hormone implants may have caused unusual weight gain in the treated mares. In 1987, after neck wounds were first observed, seven animals were darted at Stone Cabin to treat collar sores. One collar was removed, and six collars were loosened. One animal was subsequently recollared. In 1988, 28 animals were darted to treat collar problems; 19 collars were removed, 6 were adjusted, 2 horses were released without adjustment, and 1 mare died as a result of recapture. Wounds were cleaned and treated with antibiotics. Collars were removed in cases where it was judged necessary to assist recovery, and the animals were released. Since October 1988, another six tight collars have been removed from the horses at the Stone Cabin study area. Monitoring The condition of the study animals and the fit of the collars are moni- tored from the air during the four census-taking flights scheduled as part of the original study design. In 1988, BLM added an additional flight to check for collar problems in the fall after the animals gained weight over the summer. These measures were applied to all of the study areas and continued in 1989. Dr. Peck recommended increasing the additional collar monitoring to four times a year at regular intervals to detect health problems in the collared horses. In his view, the current schedule of monitoring flights is not sufficient to identify collar problems before they become severe. How- ever, several observers expressed concern that increasing the frequency of the monitoring flights, especially during the spring and summer, could increase the incidence of abortions or orphaning of foals. Other Study Areas The collar-related problems experienced at study areas other than Stone Cabin have been less serious (Table 4-1). Five animals were darted to treat
RESEARCH CONCERNS 29 complications from tight collars at Clan Alpine; no tight-collared animals were found at Flanigan or Beaty Butte. Of the five at Clan Alpine, two collars were removed, one was adjusted, and one was left in place without adjustment. Three additional collars were adjusted at Clan Alpine, and one animal died during recapture. The Flanigan and Beaty Butte study areas involved vasectomized stal- lions only. Collars became loose on several of the experimental animals. In these cases, the collars rode up over the ears of the animals, causing abrasions and concern that they might interfere with vision, hearing, or normal behavior. To correct problems with loose collars, 11 horses were darted at Flanigan through September 1988; nine collars were removed and two adjusted. During the same period of time, one horse was darted at Beaty Butte and the collar was removed. Collars also fell off the vasectomized stallions. By September 1988, 13 of 20 vasectomized males in Flanigan no longer had collars because of removal or loss. In Beaty Butte, 7 of 20 collared animals either lost collars or had them removed. Two additional collared animals died in each area during the course of the study. Since October 1988, one tight collar was removed at Clan Alpine, while a total of eight loose collars were removed at Flanigan, Clan Alpine, and Wassuks. During the fall of 1988 and throughout 1989, the BLM and the research team followed a policy of removing all problem collars, whether or not there was any injury to the animal. Behavior Effects of Collars Questions have arisen about the effects of the collars on animal behavior, because the collars may restrict sight or hearing or inhibit normal behaviors such as ear signaling, nuzzling, or neck rubbing. Dr. Cheryl Asa, an animal behavior researcher from the University of Minnesota, told the committee that her field observations of vasectomized and control stallions in the Flanigan and Beaty Butte sites indicated that the collars themselves did not seem to impede normal behavior significantly and did not compromise the animals. In fact, Dr. Asa expressed a concern that loose collars were being adjusted unnecessarily. In her view, the risk from darting the animal to adjust the collar greatly exceeded any risk from the loose collar itself. Mortality Caused by Collars The research team has attributed two deaths to collar-related problems. The first was a 25-year-old mare that died at Stone Cabin after being darted to treat a tight collar. The second loss occurred when a stallion fell off a cliff after being darted at Clan Alpine. Other animals with collars were found dead. One had a collar imbedded
30 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY in its neck, but no cause of death was determined. Another animal was found dead 12 days after she had been darted but failed to succumb. She was judged to have been dead for 3 to 7 days. The causes of both of these deaths were classified as "other." The research team classified as natural the deaths of an additional 21 collared horses (4 with marker collars and 17 with radio collars) that were found dead before August 1988. The team found no evidence that these collars were stained from drainage, as were those that had become embed- ded in the horses' necks before removal. BLM observers believe that collar- related wounds may have caused or contributed to the deaths of at least some of these animals. It is not now possible to determine if the collars played a role in any of these deaths. Effect on Statistical Adequacy Questions have arisen regarding whether deaths and collar removals re- duced the sample sizes to the degree sufficient to compromise the statistical validity of the study. A total of 273 mares and 45 stallions in the experimental populations survived roundup and initial collaring. Eighteen placebo-treated mares died at Clan Alpine shortly after being rounded up and collared. The stallions were collared first, and subsequent collar problerris with males involved loose collars that slid up over the animals' ears. By the end of the study period, 13 of 20 vasectomized animals in the Flanigan site had lost collars and 2 others had died. In Beaty Butte, 7 of 20 lost collars over the course of the study and 2 other collared animals had died. This gradual loss of collars and a lack of controls throughout the vasectomy study make the results of the vasectomy experiments difficult to interpret. Tight collars were a problem only with mares. Of the 273 collared mares in the fertility study, 23 or 8.4 percent were lost from the experimental population due to collar removal or loss. An additional 17 horses with radio collars, or 6 percent, died of natural causes during the study. Four other animals died during collaring procedures, and two more with collars were found dead but their deaths were classified as "other." It is not known whether these last six animals were wearing radio or marker collars. The number of mares from which radio and marker collars were removed and not replaced and the number of collared animals that died total 76, or 9 percent of the number originally collared. As discussed above and summarized in Tables 2-4 and 2-5, the number of animals in each treatment block both in 1988 and 1989 was sufficient to establish statistically significant differences between treatments and placebo implants in all of the cases. Regaining the animals lost to the study would not significantly strengthen the results of
RESEARCH CONCERNS 31 the research or the statistical significance of the inhibiting effect of the implants on reproduction. Summary To the extent possible, the committee has assessed the problems experi- enced with the radio and marker collars. There is no doubt that some of the collared animals suffered large and painful wounds, and at least two horses died as a direct result of problems with their collar adjustment. Twenty- three radio-collared horses were lost from the observable population because of collar removal, and two others died while being recaptured to remove their collars. Marker collars were removed from another 28 horses that were not included in the study population. No horses are known for certain to have died from collar wounds. The collar problems have been attributed to various causes, which include: the design of the collars; the material and construction of the collars; irritation caused by the radio units; tight or loose fitting collars; natural growth of young horses; rapid weight gain because of abundant forage; abnormal weight gain as a result of the hormone implants; and difficulty in making fine adjustments at the first collar fitting. Most observers agree that most of the neck injuries were related to the radio units. The collars were based on a generally accepted design. Similar collars have been used successfully in field studies of wild and free-roaming horses and many other species of large animals. The research team has had extensive experience with collars, telemetry, and fieldwork of this nature. Based on prior experience, the problems with collars experienced in this study were not anticipated. Both the research team and the BLM responded as collar problems were observed. Animals with collar wounds were recaptured, and the wounds were treated and the collars removed as necessary. Treated injuries have healed, and some of the collars have been replaced. Additional flights were made to locate horses with collar injuries, and BLM added a fall flight to the monitoring schedule to check the condition of the horses after the period of summer weight gain. Problems with collar injuries could continue during the remaining months of this study because the construction of the collars, the radio units, and the potential for weight gain remain unchanged. However, because the research team and BLM are now more intensively monitoring the problem, future injuries should be reduced in both incidence and severity. The BLM will round up all collared animals after the 1990 observation season and remove the collars.
32 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY FOAL ORPHANING AND LOSS Questions have arisen over whether the spring and summer monitoring flights, to assess the reproductive status of the mares, separate mares from their foals and cause foal orphaning or death. In the case of collared mares, the numbers on the collars must be read in order to determine whether they are steroid-implanted experimentals, placebo implants, or untreated mares that have been fitted with marker collars. In order to read the numbers, the helicopter must descend to a low altitude over the mares, and they must at times be chased some distance before the numbers can be read. The con- cern is that, in this monitoring process, mares and foals can be separated and may not be reunited. In such cases, the foals might be permanently separated from the mares at ages too young to survive on their own. Prior to the committee's 1988 meeting in Reno, Nevada, the committee conveyed concerns to the research team arising from aerial monitoring and the associated risk of foal orphaning and death. At the meeting, the committee examined two sources of evidence provided by the research team to determine whether foal orphaning or loss was occurring. The first source was the record of observations on collared mares to determine whether mares seen with foals in the initial spring flights were observed without young in subsequent flights. This record proved inconclusive because of ambiguous trends. For example, a mare that may have been seen with her foal in April and sighted twice in May without her foal may have been seen in June with her young. Because of the frequency of these cases, no general pattern of individual mares could be inferred. The second source of evidence consisted of aggregating all of the flight observations and calculating the percentage of mares with foals seen in each study area. A decline through the four flight periods would suggest foal loss. Trends were calculated for Wassuks and Stone Cabin in 1987, and Wassuks, Stone Cabin, and Clan Alpine in 1988. Here again, the results were incon- clusive. Because the samples were small and individual flight results were therefore variable, trends were difficult to infer. Two areas showed a decline: Wassuks in 1987 between censuses 3 and 4, and Clan Alpine in 1988 between censuses 2 and 4. Stone Cabin in 1988, however, showed a slight, net increase between censuses 1 and 4. Stone Cabin in 1987 and Clan Alpine in 1988, although variable between censuses, showed no net trend between censuses 1 and 4. In light of available evidence, the committee concluded that the monitoring did not appear to have caused heavy or consistent foal loss. This conclusion did not rule out the possibility that individual foals may have been lost. If such a loss occurred, however, it was not frequent enough to detect through aerial observations. In addition, some natural foal mortality is expected.
RESEARCH CONCERNS 33 Earlier studies have shown that the foal stage is the period of highest natu- ral mortality in the life of a wild horse. The issue arose again after the 1989 field season, sparked by a memorandum (Sweeney, 1989) that was based on field observations by BLM employees who accompanied the research flights, in a separate helicopter, to identify collared mares and to observe foals. They recorded instances during 1989 when the BLM observers feared that foals may have been separated. After 17 flights over the Wassuks and Clan Alpine areas, one observer wrote, "In my opinion, at least five foals were likely permanently separated from their mares. ... I believe that the [actual] number of foals [lost] ... is greater than the five." Sweeney's memorandum concludes "an undetermined num- ber are left behind to become orphaned." The committee again raised the matter with the research group at the February 1990 review in Denver. In response, the researchers described the flight procedures as follows: â¢ No band is chased more than 400 to 500 yards. If a foal separates from a band, the research helicopter pulls out of the chase. â¢ If a foal is separated, the helicopter circles in front of the band and turns it back toward the foal. â¢ As a precautionary measure, all research flights are accompanied by a second helicopter containing BLM observers who are watching for excessive chases and foal separation. If these occur, they can radio the research helicopter and advise it to end the chase. BLM has never called off the research helicopter. As in the 1988 review, the research group also reviewed the trends in percentages of observed mares with foals over the four successive flights in the three study areas in 1989. Here again, there were no consistent trends. A review of available data did not support the assertion that a major, consistent loss of foals occurred. Some small number of foals could have been lost, although even this is not unequivocally shown. The committee remains concerned about the possibility that the monitor- ing flights may be causing foal orphaning or loss, and urges that the research team proceed with caution. From the standpoint of the research protocol, both placebo- and steroid-implanted mares are equally at risk. Thus, to the extent there may have been some losses, a comparison of the fecundity of the steroid- and placebo-implanted mares is not invalidated. ABORTION Another concern is the possibility that roundup for treatment and subse- quent-year censusing flights cause mares to abort. Some observers maintain that (1) abortion may be caused by the stress the animals endure under the
34 WILD HORSES: FIELD STUDIES IN GENETICS AND FERTILITY research methods of roundup and censusing, and (2) abortion, if it occurs, might have a statistical impact on the research results with regard to the actual effectiveness of the contraceptive implants. Because of observed abortion in corrals following BLM herd-control roundups, an analogy has been drawn between these observations and the possible occurrence of abortion as a result of research-induced stress. The research team has seen no fetuses or other evidence of abortion. Of course, this does not rule out the possibility of its occurrence. It simply may not have been seen. However, as discussed above and in Table 2-4, the foaling rate of the Clan Alpine placebo-implanted mares in 1988 (42 percent) and that of the steroid-implanted mares (53 percent) were similar to the foaling rates of the untreated, marker-collared animals that were not approached closely by he- licopter (47 percent). The Wassuks placebo-implanted mares also had a similar foaling rate (45 percent). All of these rates fall well within the 95 percent confidence intervals of each other. These results do not entirely rule out the possibility of abortion. Relatively small samples and year-to-year and site-to-site variation could mask the occurrence of some abortions. However, given the similar foaling rates, if abortion is occurring, it is occurring at too low a rate to be measured statistically. With regard to the integrity of the research methods and the validity of the results, steroid- and placebo-implanted mares are subject to the same conditions in the conduct of the research. However, because steroid-treated mares are expected to have fewer pregnancies, more abortions would be expected from the placebo-implanted mares. Therefore, abortion, if occur- ring, would narrow the difference between the foaling rate of the steroid group and that of the placebo group. Because the foaling rate of the placebo-implanted mares in the Clan Alpine area was within the range of the comparative values of less dis- turbed, marker-collared mares in 1988, there is no reason to conclude that abortion resulted in a sufficiently narrowed difference between steroid and placebo groups to invalidate the experiment. In the absence of either (1) direct evidence or (2) a statistical indication of abortion resulting from study protocols, the committee concludes that the validity of comparing foaling rales in steroid- and placebo-implanted mares is not compromised by abortion, if any occurs. DISAPPEARANCE OF PENNED ANIMALS AT LOVELOCK CORRALS The claim has also been made that large numbers of experimentally treated animals disappeared in the Lovelock corrals. According to the researchers, however, the total number of missing animals was 13. The number of
RESEARCH CONCERNS 35 animals in this treatment group was 210 in the first series, 70 in the second, and 50 in the third. These groups were not mutually exclusive; some animals in the first series were used again in the second and third. Hence, the total number of experimental mares was something less than the sum of these three numbers. Nevertheless, the sample sizes were substantial and the disappearance of 13 animals does not pose a serious problem for the integ- rity of the experiments. FOOD CHAIN RISKS FROM STEROIDS Questions have been raised regarding the wisdom of releasing into the wild large numbers of mares with steroid dosages coursing through their tissues. These animals might eventually be rounded up and used for human consumption, or they might die and their flesh eaten by wild carnivores whose reproduction could be impaired. The dosage trials in the Lovelock corrals were undertaken with estradiol, a natural estrogen that is chemically broken down by the digestive process if consumed by an animal. The first dosages administered were not effective. In the second and third series the estradiol dosages showed some effective- ness, but time was running out and an effective dosage was needed to treat mares in the field by September 1986, according to the schedule of the contract. Therefore, pen trails were begun with ethinylestradiol, a synthetic estrogen that is 30 times as effective as natural estradiol, but withstands digestion if consumed by an animal. This treatment proved effective both in pen- and field-treated animals. As to the risk of consumption, the possibility of a treated animal being eaten is very small. Moreover, ethinylestradiol is the steroid commonly used in human oral contraceptives. Using liberal calculations, a human must consume 1 pound of horse flesh shortly after implantation (i.e., the peak level of steroid release) in order to ingest the amount of ethinylestradiol contained in a single, low-dose oral contraceptive pill. At some time after implantation, when the steroid release of the capsule had declined to its longer-term level, the amount ingested would be substantially less.