Northern Prairie Wildlife Research Center
In Part III, we provide recommendations for interpreting evidence of depredation found at duck nests and apply them to duck nests found on waterfowl production areas in Minnesota, North Dakota, and South Dakota.
Identification of predators responsible for destruction of duck nests involves interpretation of circumstantial evidence. Thus, results are always subjective rather than definitive. Our examination of literature and our research suggest potential for error in interpreting evidence of depredation. We question the accuracy of identification of offending predators in most literature because of limited information available to investigators attempting to make determinations. Investigators not treating all principal predator species present in their study areas as potential depredators of nests, or investigators unaware of variations within or similarities among depredation patterns of certain predators, were likely to have made identification errors. We hope our findings will cause investigators to be more cautious in identifying depredators of duck nests, and to base their findings on objective rather than subjective methods.
Investigators interested in identifying predators of duck nests can use information provided herein as a field guide while examining evidence at nests. However, investigators interested in credible data for research or management should not rely on identifications made at nests. Rather, they should carefully record evidence of depredation on a nest depredation record (e.g., Appendix D, Fig. 1) for subsequent determination and analyses.
Our findings suggest much variability in evidence left at nests by individuals of some predator species and much similarity in evidence left by individuals of different predator species. Moreover, other factors (e.g., age of nest, weather, habitat) can affect amount and types of evidence of depredation found at nests. These variants, however, do not preclude usefulness of recording evidence of depredation.
We found major differences in depredation patterns among most predator species, with some species being easier to identify from evidence found at nests than others. Moreover, we found that certain factors (e.g., knowledge of the predator community, restricting analyses to certain types of nests) often can be used to improve the chances of identifying offending predator species. By using all available information, it often is possible to estimate the proportion of nests destroyed by a particular predator species.
Verification data for individual predator species generally are scant and based on small samples, as are data concerning frequency with which most predators leave various types of evidence. We found treatment of eggs, types of damage to shells, and location of openings in eggshells to be most useful for distinguishing among depredations by various predator species. Other evidence enables ruling out certain predator species because of its uniqueness to 1 or a few species (e.g., cached egg, dug area, dead hen, conspicuous yolk residue), or to substantiate other evidence (e.g., locations of eggshells, displacement of nest material).
Investigators should use judgement when applying our findings. For example, our findings for artificial nests probably exaggerate amount of nest material likely to be displaced by predators, because nest material at natural nests is packed more tightly. Also, hens at natural nests may return displaced nest material. However, our findings that some species displaced little or no nest material from loosely packed artificial nests (e.g., raccoon, gulls) is strong evidence that these predators are unlikely to displace nest material from natural nests.
We limited our recording of evidence of depredation to factors that could be accurately and easily recorded by investigators with varied backgrounds. Investigators, however, may wish to record other evidence that they are confident pertains to a predator species. Such evidence might include trails and matted areas at nests, tooth puncture marks in eggshells, shapes of openings in eggshells, and edges of openings in eggshells (e.g., finely serrated, connected pieces caved inward).
There are 3 steps to estimate the proportion of nests destroyed by a predator species: (1) defining the question, (2) delineating the data set, and (3) establishing a hierarchy of criteria for assigning a destroyed nest to a predator species. Details of each step will vary depending on investigator needs and amount and quality of available data.
Defining questions to be addressed is a critical first step. A question, such as - What is the estimated proportion of the nests destroyed by red foxes? - is too general. For evaluations to proceed, that question must be refined using qualifiers such as duck species, time periods, habitat, and age and size of clutch. A precise question might be: What is the estimated proportion of nests of dabbling ducks initiated during May and June in uplands and have ≥6 eggs at time of destruction that are destroyed by red foxes?
After the question has been defined, the data base is delineated using information from both nest records and nests depredation records. We recommend that evaluations be restricted to nests with ≥6 eggs when last visited by an investigator before the nest was destroyed. These nests contain a sufficient number of eggs for a predator to establish its pattern of depredation. Moreover, these nests are unlikely to be abandoned by the hen because of investigator disturbance, as often occurs with nests discovered early in the egg laying phase (pers. observations). It may be advantageous to restrict evaluations to nests with incubated eggs when last visited by the investigator, because additional eggs are rarely added to such nests. This enables determining with greater certainty the proportion of eggs in the clutch at the time of destruction that are represented by eggshells of the various types.
Investigators also may wish to restrict evaluations to nests discovered destroyed within a specific time interval after the nest was last visited, because evidence of depredation deteriorates and/or becomes compromised over time. During 1992, we revisited destroyed duck nests at 11 locations in South Dakota to determine temporal changes in occurrence and appearance of eggshells. Data were from 81 nests with ≥1 eggshell at the nest when discovered destroyed, and no additional eggs depredated between subsequent visits. Evidence of depredation was not disturbed by investigators. Each nest was revisited on average 8.9 days (range = 2-16 days) after the nest was discovered destroyed, for a total of 718 exposure-days. Of 393 eggshells present during the first visit, 348 (89%) remained during the second visit for an average loss rate of 0.06 eggshells/exposure- day. There was no change in number of eggshells at 37 (46%) nests. All nests with shell fragments during the first visit (n = 75) had shell fragments during the second visit. We also examined fate of whole eggs (n = 101) found at 18 nests during the first visit. Additional eggs (n = 83) were depredated at 15 (83%) of these nests during the average 8.1-day revisit period. Based on these findings, we recommend restricting analyses to nests destroyed during a revisit interval of ≤2 weeks.
The purpose of the hierarchy of criteria is to reduce the overall data set to nest depredation records with evidence characteristic of the predator species of interest. This is done by first excluding records with evidence unlikely to have been left by that species. For remaining records, criteria characteristic of depredation by species of interest are ordered from most to least definitive, with the most definitive treated first. At any step, records with evidence unique to that species (e.g., cached eggs for badgers) can be assigned to the species, and the query continued to isolate records with other evidence characteristic of the species. With each subsequent query, number of candidate records in the pool decreases until no further exclusions or retentions are desired or possible. The proportion of initial records remaining after queries are completed, plus those already assigned to the species, is the total proportion of destroyed nests attributed to the predator species.
In general, the likelihood of obtaining definitive estimates of the proportion of nests destroyed by a predator species increases as number of predator species decreases. Investigators should establish criteria for assigning destroyed nests to predator species only from information for species that were present in habitats where nests were located. Because the suite of predator species at individual sites varies greatly throughout the Prairie Pothole Region (Sargeant et al. 1993), we urge investigators to obtain predator population data as part of their studies of duck nest success.
Sovada et al. (1995) used the above approach to assign destructions of duck nests to red foxes in their evaluation of differential effects of coyotes and red foxes on duck nest success. The suite of predator species at each study area was determined and central to the evaluation. They found that the proportion of nests of ≥6 eggs depredated in a manner attributed to red foxes was lower in the coyote areas (2%) than in the red fox areas (17%). The findings confirmed the authors' hypothesis that the difference in nest success (32% [coyote areas] vs. 17% [fox areas]) between areas occupied by each species was attributable to the difference in the canid community.
We used combined data from Sovada et al. (1995) and from sites without predator removal (Sargeant et al. 1995) to provide examples of interpreting data recorded on nest depredation records. We examined depredations by red foxes, striped skunks, and raccoons (Appendix B, Table 7,Appendix B, Table 8, Appendix B, Table 9). These species are considered to be major predators of duck nests in the Prairie Pothole Region (Sargeant and Arnold 1984).
Data were from destroyed nests in uplands of 48 federal waterfowl production areas in Minnesota, North Dakota, and South Dakota during 1987-92. Data collection was similar in both studies. Nests were found throughout the duck nesting season and visited at about 7-10 day intervals until fate was determined. Data from each nest were recorded similar to procedures described by Klett et al. (1986). Evidence of depredation was recorded on prototypes of the nest depredation record (Appendix D, Fig. 1).
Predator population data revealed that striped skunks and raccoons were common at all areas. Coyotes, red foxes, and Franklin's ground squirrels were common at about half of areas. American badgers, minks, and weasels were present at most areas but generally were not common. American crows, black-billed magpies, and gulls were absent or only occasional visitors at nearly all areas. Based on these findings, we excluded avian species as potential depredators of the duck nests. We wished to estimate the proportion of nests that were destroyed by red foxes, striped skunks, and raccoons. Qualifiers for delineating the data set included the following: (1) nests of all duck species in all months, (2) nests in uplands only, (3) nests of ≥6 incubated eggs when last visited before being destroyed, and (4) nests discovered destroyed ≤2 weeks after the nest was last visited by an investigator (Appendix B, Table 7,Appendix B, Table 8,Appendix B, Table 9). Qualifiers 1 and 2 resulted in a sample of 636 destroyed nests. Qualifiers 3 and 4 and removal of 6 other nests with incomplete information reduced the sample for analyses to 389 nests.
For foxes, we selected nests with no eggshells or shell fragments at the nest (Appendix B, Table 7). These criteria are strongest indicators of nest depredation by red foxes. These criteria resulted in retention of 86 (22%) nests. That sample was further reduced to 81 (21 %) nests by eliminating nests with dug areas, carcasses or carcass parts, or evidence of a cached egg at the nest. Such evidence is rarely left by red foxes. Some investigators may further reduce the sample by eliminating nests with ≥l whole egg at the nest and nests with >30% of nest material displaced (evidence seldom left by red foxes at nests). Thus, we estimated that 19-21% of the nests were destroyed by red foxes.
For striped skunks and raccoons (Appendix B, Table 8,Appendix B, Table 9), we first excluded from the usable sample of 389 destroyed nests those with evidence indicating the hen was killed or an egg was cached. Both species rarely leave such evidence. The hierarchy of nest depredation criteria for the 2 species then diverged.
For striped skunks, we retained only nests (n = 118) with eggshells of >50% of eggs present when the nest was last visited by an investigator. This is a strong indicator of depredation by both striped skunks and raccoons because both species seldom remove eggs from nest sites when they eat the eggs. We then eliminated nests for which >50% of eggshells had small holes, those with >50% of openings in eggshells in an end of the eggshells, and those with >25% of eggshells >1 in from the nest. These are strong indicators that a nest was not depredated by skunks. Application of the above criteria reduced the sample of usable nests to 87 (22%). A conservative estimate of 69 (18%) nests destroyed by skunks was obtained by eliminating nests with ≥1 dug area at the nest, those with ≥1 whole egg at the nest, those with no eggshell fragments at the nest, and those with no nest material displaced. These types of evidence are only occasionally found at duck nests with clutches destroyed by striped skunks.
For raccoons, we eliminated nests with ≥1 dug area at the nest, a behavior never exhibited by raccoons in our study, and then retained nests with eggshells of >50% of eggs present when the nest was last visited. These criteria reduced the sample to 110 (28%) nests. We then eliminated nests for which >50% of the eggshells had small holes and those with >50% of openings in eggshells in the side or a side-end of the eggshells. These are strong indicators that a nest was not depredated by raccoons. This process left 31 (8%) nests likely destroyed by raccoons. A conservative estimate of 17 (4%) of the usable nests destroyed by raccoons was obtained by eliminating nests with >25% of eggshells >1 m from the nest, those with >30% of nest material displaced, those with ≥1 whole egg at the nest, and those with no eggshell fragments at the nest. These types of evidence are only occasionally found at duck nests destroyed by raccoons.
In these examples, we identified the offending predator for about 40% of the usable nests. We found that red foxes and striped skunks were major depredators of the nests, and that raccoons were minor depredators of the nests. About 60% of nests were not assigned to the 3 examined species. Most of those nests probably were destroyed by other species. Some may have been destroyed by these 3 species, but were unassigned because of variations in the manner in which individuals destroyed nests. Others may have been sequentially visited by ≥2 species, each leaving additional evidence of depredation.
Using the above hierarchy of criteria, it is possible to assign destruction of an individual nest to >1 predator species. In our examples, no nest with destruction assigned to red foxes was assigned to the other species. However, this would have been likely had our evaluation included destructions by other species such as American crows and Franklin's ground squirrels. For raccoons and striped skunks, 11 nests were assigned to both species. This is to be expected because of similarities of depredation patterns of certain individuals of these species.
The above examples illustrate how data recorded on nest depredation records can be interpreted to assign destroyed nests to predator species and provide results suitable for comparison among studies. Other investigators might have chosen slightly different sets of qualifiers and/or slightly different hierarchies of criteria to estimate proportion of nests destroyed by each species and obtained slightly different, but equally valid, results. As more information on nest depredation patterns of individual predator species becomes available, ability of investigators to more completely assign destructions of nests to predator species will improve.