Limited Capacity of Deer To Serve as Zooprophylactic Hosts for Borrelia burgdorferi in the Northeastern United States

ABSTRACT

Because deer are considered to be incompetent reservoirs of the agent of Lyme disease (Borrelia burgdorferi sensu stricto) in the northeastern United States, they may serve as zooprophylactic or “dilution” hosts if larvae of the deer tick vector (Ixodes dammini, “northern” clade of Ixodes scapularis) frequently feed on them. To determine whether host-seeking nymphal deer ticks commonly feed on deer as larvae, we used a real-time PCR host bloodmeal remnant identification assay to identify the host on which these ticks had fed. Nymphal lone star ticks (Amblyomma americanum) were collected simultaneously in our sites and provided an index of the availability of deer in these sites. At 3 of the 4 sites, Ixodes nymphs had fed as larvae on a variety of hosts, including mice, birds, and shrews, but rarely on deer (<6% for all sites); in contrast, lone star tick nymphs had commonly fed on deer (31 to 78%). Deer were common larval hosts for Ixodes ticks (39% of bloodmeals) in only one site. The prevalence of B. burgdorferi in host-seeking nymphal deer ticks was associated with mouse-fed ticks (P = 0.007), but there was no association with deer-fed ticks (P = 0.5). The diversity and prevalence of hosts that were identified differed between deer ticks and lone star ticks that were collected simultaneously, demonstrating that there is no confounding of host bloodmeal identification by contaminating environmental DNA (eDNA). We conclude that deer were not common hosts for larval deer ticks, thus limiting their zooprophylactic role in our sites.

IMPORTANCE Because deer are incompetent reservoirs for B. burgdorferi, their presence may modulate the force of enzootic transmission by serving as zooprophylactic or “dilution” hosts. Such an effect would depend on the extent to which subadult deer ticks feed on other hosts. We used bloodmeal analysis on nymphal deer ticks to identify the host upon which larvae had fed. We found that lone star ticks collected at the same time as deer ticks commonly fed on deer, but deer ticks did not. We conclude that deer are not a preferred host for larval deer ticks and, thus, are not necessarily zooprophylactic.

INTRODUCTION

Perpetuation of the agent of Lyme disease (Borrelia burgdorferi sensu stricto) in the northeastern United States depends on its deer tick vector, Ixodes dammini (the “northern” clade of Ixodes scapularis), focusing its bites on reservoir competent hosts, such as white footed mice, chipmunks, and shrews. The abundance of deer ticks is causally associated with that of deer (1, 2), but these hosts are incompetent B. burgdorferi reservoirs (3). The extent to which deer feed subadult deer ticks has been thought to reduce the local force of enzootic B. burgdorferi transmission; deer are zooprophylactic or “dilution” hosts (3). Although conceptually straightforward, the effect of zooprophylaxis or “dilution effect” on B. burgdorferi transmission varies as a function of local vertebrate diversity and the availability of alternate hosts. The use of deer exclosures demonstrated that the force of B. burgdorferi transmission (as measured by the prevalence of infected host seeking ticks) was greater in plots without deer (4, 5), but more long-term analyses suggest that the effect is more complex (6, 7); model-based analyses tend to support the zooprophylaxis effect (8–13). It is likely that the zooprophylactic effect is dynamic, with an initial increase in transmission due to the increased number of ticks and a subsequent decrease in transmission when the deer population reaches a certain size and starts to divert subadult tick feeding from reservoir competent hosts (14).

Deer ticks have little host specificity, although adult ticks do not feed on small mammals or birds; infestation of a host depends on passive encounter rates. Encounter rates are directly related to population densities and diversity of the species that comprise the local community (15). Zooprophylaxis would not occur if ticks had strong host preferences and certain preferred hosts fed a disproportionate number of ticks. The main evidence for zooprophylaxis as a modulator of the force of enzootic B. burgdorferi transmission is indirect and comes from extrapolations of observed infestation of diverse trapped animals in sites and from models (11, 16). We can now directly identify the animal from which a host-seeking tick obtained its bloodmeal in the prior developmental stage using a highly sensitive PCR assay that targets mammalian retrotransposons (17). We found that the proportion of deer tick nymphs that had fed on deer as larvae varied greatly depending on the year and the site, ranging from 14% to 75% from 4 different field sites (17). This was unexpected because deer abundance tends to be relatively stable from year to year (strong density dependence of recruitment (18) in a stable habitat; our sites are not disturbed or overexploited for hunting). Another unexpected finding was evidence of multiple bloodmeals, particularly of individual ticks feeding on both deer and mice and indeed as many as 36% of ticks at one study site. Other host bloodmeal identification studies using other methods also detected evidence of multiple bloodmeals, ranging from 10% to 23% of ticks for which there was successful host identification (19–22). The finding of host-seeking ticks with evidence of having fed on more than one host is not particularly consistent with our current paradigm of how ticks feed; we assume that ticks attach to a single host and feed successfully to repletion once attached. Partial bloodmeals would occur only when a tick is groomed off or the host dies (23), but the frequency with which this occurs in nature is unknown. Partial feeding and reattachment may have implications for the maintenance of B. burgdorferi diversity in a site as well as for quantitative models of transmission.

Because subadult ticks (other than one-host ticks) spend most of their lives on the ground, they are likely to have contact with environmental DNA (eDNA) from the community of animals living in the site. eDNA has been shown to confound the results from other types of experiments that rely on highly sensitive PCR methods, such as describing the microbiome of ticks (24, 25). It may be that the results of bloodmeal identification are also being similarly influenced, and the multiple bloodmeals that we and others have detected are simply due to contaminating mammalian DNA from the environment. Accordingly, to determine whether eDNA contaminates host-seeking ticks and thereby confounds bloodmeal analyses, we analyzed samples of two species of ticks (deer ticks and lone star ticks [Amblyomma americanum]) with different expected host preferences that were collected from the same habitat at the same time. Because larval lone star ticks rarely feed on small rodents, such as Peromyscus leucopus (26, 27), the finding of P. leucopus DNA from host-seeking nymphal lone star ticks would suggest the general degree of mouse eDNA contamination within the site. In addition, because deer are heavily infested by all stages of lone star ticks (reviewed in reference 28), the prevalence of host-seeking nymphal lone star ticks that had fed on deer serves as an index of the availability of deer in the study site, thereby allowing for a test of the hypothesis of deer zooprophylaxis on the prevalence of B. burgdorferi infecting host-seeking nymphal deer ticks in the same site.

RESULTS

In preliminary experiments, we determined that washing nymphal ticks with 5% bleach before nucleic acid extraction was sufficient to remove external contamination that had been deliberately applied. All of the lab-reared ticks that had been incubated with deer hair that were not externally decontaminated tested positive for both deer and mouse, but all of the ticks that were decontaminated with bleach tested positive for mouse only (data not shown).

There was no difference in the number of multiple bloodmeals between deer tick nymphs from Nantucket that were externally decontaminated with bleach (n = 87) compared to those that were not (n = 89) (Table 1). However, externally decontaminated ticks comprised significantly fewer mouse-fed ticks (9.2% bleached versus 22% no bleach; P = 0.02) and more ticks for which we could not identify a bloodmeal source (34% bleached versus 20% no bleach; P = 0.04), indicating that, at least for our Nantucket study site, many host-seeking nymphs were externally contaminated with mouse DNA.

TABLE 1.

Bloodmeal identification of deer tick nymphs from Nantucket comparing ticks externally decontaminated by bleach to those that were not^a

Host	No. bleached (%)	No. unbleached (%)	Significance (P value)b
Multic	1 (1.1)	4 (4.4)	ns
Bird	4 (4.5)	9 (10)	ns
Deer	34 (39)	32 (36)	ns
Mouse	8 (9.2)	20 (22)	0.02
Shrew	12 (14)	14 (16)	ns
Vole	0	0	nd
Squirrel	0	0	nd
Rabbit	0	0	nd
Unknown	30 (34)	18 (20)	0.04
Total	87	89

^aNote that the squirrel primers amplify other Sciuridae, but only squirrels are present on Nantucket.

^bnd, not determined; ns, not significant.

^cMulti, ticks that tested positive for multiple bloodmeal hosts.

Ticks were sampled from Robin’s Island and Prudence Island inasmuch as there are no lone star tick infestations in our Nantucket Island study sites. On Robin’s Island, the deer tick nymphs had commonly fed on mice as larvae (79%), but lone star ticks rarely did (5.9%) (Table 2). By contrast, lone star tick nymphs had commonly fed on deer as larvae (78%), but deer ticks rarely did (5.1%). External decontamination of lone star ticks by bleach did not change these results (Table 3). Unfortunately, there were insufficient samples of deer tick nymphs collected from this site to do a similar comparison, and thus we sampled two sites on Prudence Island where both commonly cooccur. In one site (Schoolhouse), a greater proportion of deer tick nymphs had external contamination with mouse DNA (52% unbleached versus 15% bleached; P = 0.004) as well as from multiple hosts (30% unbleached versus 6% bleached; P = 0.02). However, there was no difference in our detection of external contamination between bleached or unbleached lone star tick nymphs (Table 4). At the other site (Blueberry Hill), there was little difference in external contamination between the bleached and the unbleached deer tick nymphs. However, the proportion of these ticks with evidence of multiple bloodmeals tended to differ (11% unbleached versus 0 bleached; P = 0.06). Unbleached lone star tick nymphs from this site, however, were much more likely to test positive for deer than those that were bleached (97% versus 46%; P = 0.000), and there were fewer ticks not yielding a bloodmeal identification (0 versus 29%; P = 0.002) (Table 5). We conclude that the prevalence of external contamination differs between the two kinds of ticks within the same collection site and that there are differences between the sites with respect to the presence and kind of contaminating DNA.

TABLE 2.

Bloodmeal identification of nymphal deer ticks and lone star ticks collected simultaneously on Robin’s Island^a

Host	No. of deer ticks (%)	No. of lone star ticks (%)	Significance (P value)b
Multic	2 (5.1)	2 (3.9)	ns
Bird	1 (2.6)	4 (7.8)	ns
Deer	2 (5.1)	40 (78)	0.000
Mouse	31 (79)	3 (5.9)	0.000
Shrew	0	0	nd
Vole	0	0	nd
Squirrel	0	0	nd
Rabbit	0	0	nd
Unknown	7 (18)	6 (12)	ns
Total	39	51

^aTicks were not externally decontaminated. Note that the squirrel primers amplify other Sciuridae.

^bnd, not determined; ns, not significant.

^cMulti, ticks that tested positive for multiple bloodmeal hosts.

TABLE 3.

Bloodmeal identification of lone star nymphs from Robin’s Island comparing ticks that were externally decontaminated with bleach to those that were not^a

Host	No. bleached (%)	No. unbleached (%)	Significance (P value)b
Multic	0	2 (3.9)	ns
Bird	2 (5.6)	4 (7.8)	ns
Deer	32 (89)	40 (78)	ns
Mouse	0	3 (5.9)	ns
Shrew	0	0	nd
Vole	0	0	nd
Squirrel	0	0	nd
Rabbit	0	0	nd
Unknown	2 (5.6)	6 (12)	ns
Total	36	51

^aNote that the squirrel primers amplify other Sciuridae.

^bnd, not determined; ns, not significant.

^cMulti, ticks that tested positive for multiple bloodmeal hosts.

TABLE 4.

Bloodmeal identification of nymphal deer ticks and lone star ticks collected from the Schoolhouse site on Prudence Island comparing ticks externally decontaminated with bleach to those that were not^a

Host	Deer ticks			Lone star ticks
	No. bleached (%)	No. unbleached (%)	Significance (P value) b	No. bleached (%)	No. unbleached (%)	Significance (P value) b
Multic	2 (6)	10 (30)	0.02	0	1 (2.6)	ns
Bird	22 (67)	18 (55)	ns	11 (48)	18 (46)	ns
Deer	2 (6)	3 (9)	ns	8 (35)	12 (31)	ns
Mouse	5 (15)	17 (52)	0.004	0	0	ns
Shrew	0	0	nd	0	0	nd
Vole	0	0	nd	0	0	nd
Squirrel	0	0	ns	0	1 (2.6)	ns
Rabbit	0	0	ns	1 (4.3)	0	ns
Unknown	6 (18)	3 (9)	ns	3 (13)	5 (13)	ns
Total	33	33		23	39

^aNote that the squirrel primers amplify other Sciuridae.

^bnd, not determined; ns, not significant.

^cMulti, ticks that tested positive for multiple bloodmeal hosts.

TABLE 5.

Bloodmeal identification of nymphal deer ticks and lone star ticks collected from the Blueberry Hill site on Prudence Island^a

Host	Deer ticks			Lone star ticks
	No. bleached (%)	No. unbleached (%)	Significance (P value) b	No. bleached (%)	No. unbleached (%)	Significance (P value) b
Multic	0	4 (11)	0.06	1 (3.6)	1 (3.4)	ns
Bird	13 (33)	11 (29)	ns	6 (21)	2 (6.8)	ns
Deer	0	2 (5.2)	ns	13 (46)	28 (97)	0.000
Mouse	17 (44)	18 (46)	ns	0	0	ns
Shrew	0	0	nd	0	0	nd
Vole	0	0	nd	0	0	nd
Squirrel	0	0	nd	0	0	nd
Rabbit	0	0	nd	0	0	nd
Unknown	9 (23)	11(29)	ns	8 (29)	0	0.002
Total	39	38		28	29

^aTicks externally decontaminated with bleach were compared to those that were not. Note that the squirrel primers amplify other Sciuridae.

^bnd, not determined; ns, not significant.

^cMulti, ticks that tested positive for multiple bloodmeal hosts.

The contribution of the individual animal species that serve as a host for ticks was site specific even from closely situated sites from the same island. Mice were the primary host for deer tick larvae only on Robin’s Island (79%); deer and birds were only marginally utilized there. On Nantucket, however, deer appeared to feed more larval deer ticks than did mice. Shrews fed a significant number as well (14%) on Nantucket, whereas birds only fed a few. On Prudence Island, birds and mice were the primary hosts for deer ticks (Tables 4 and 5), with birds being more important at the Schoolhouse site than at the Blueberry Hill site. The prevalence of B. burgdorferi varied from 20% on Nantucket to 59% on Robin’s Island (Table 6) and was positively associated with the proportion of ticks that fed on mice (R² = 0.99; P = 0.007); there was no association with the proportion of ticks that fed on deer (R² = 0.25; P = 0.5) (Fig. 1). For lone star tick larvae, deer were generally the most important host. On Robin’s Island, lone star tick nymphs had fed as larvae primarily on deer (89% of the bleached ticks). However, on Prudence Island, deer were the primary host only at the Blueberry Hill site. At the Schoolhouse site, birds fed more lone star tick larvae than did deer, but this was not statistically significant (48% bird versus 35% deer; P = 0.5).

TABLE 6.

Estimated prevalence of B. burgdorferi in deer tick nymphs from each site

Site	No. tested	No. positive	Prevalence estimate (%) (95% confidence interval)
Nantucket	179	36	20.1 (14, 27)
Robin’s Island	39	23	59.0 (42, 74)
Prudence Island, Schoolhouse	66	14	21.2 (12, 33)
Prudence Island, Blueberry Hill	77	27	35.1 (25, 48)

FIG 1.

Correlation between the prevalence of B. burgdorferi in deer ticks to the percent of ticks that had fed on either mice or deer. The estimated prevalence of infection for B. burgdorferi for each site was plotted against the percentage of ticks that had fed on either mice (black dots) or deer (red squares) and linear regression lines were drawn. The prevalence of B. burgdorferi was significantly associated with ticks that had fed on mice (R² = 0.98; P = 0.007) but not with ticks that had fed on deer (R² = 0.25; P = 0.5).

DISCUSSION

We consistently detected differences in bloodmeal identification from ticks that had been externally decontaminated compared to those that had not. However, the contamination does not appear to arise from eDNA that is likely found at our field sites, nor does it appear to be introduced by our handling of the ticks, e.g., from our well-used tick drags or forceps. We would expect eDNA to reflect all animals present in a habitat, with more contamination deriving from the larger animals that leave a greater “footprint” in the environment, that is, more shed hair, scat, or urine. Thus, we would expect that deer would be the most abundant source of contaminating eDNA from most of our sites. Furthermore, such general contamination would be similar for all ticks collected from the same site at the same time. Any contamination derived from our handling would also be similar for all ticks from the same site since they were collected simultaneously and handled identically. This is not what we found. Deer tick nymphs from Nantucket and Prudence Island demonstrated evidence of external contamination from mice, but not other hosts, such as deer, birds, or shrews. Most of the multiple bloodmeals identified from host-seeking ticks from those sites also were positive for mouse DNA; we, thus, cannot assume that the detection of mouse DNA from a host-seeking tick with evidence of multiple bloodmeals as a larva automatically implies that the larva had fed on a mouse and that evidence of other hosts comprise eDNA contamination. The lone star tick nymphs from the same sites, collected on the same drags with deer tick nymphs, were not similarly contaminated. Instead, most of the lone star tick nymphs demonstrated no external contamination at all. We could only identify contaminating deer DNA on lone star tick samples from a single site. We conclude that contaminating DNA on ticks was not from the general environment (eDNA mixture), but instead, was from a specific host. Furthermore, deer tick and lone star tick nymphs cooccurring in the same leaf litter and vegetation differ in the degree of their external contamination, again arguing against a major influence of eDNA on our host bloodmeal identification assay.

External decontamination significantly reduced the number of multiple bloodmeals detected from host seeking nymphal deer ticks, but they were not eliminated, demonstrating that multiple bloodmeals do indeed occur, albeit at a frequency less than what we previously estimated (17). Evidence for multiple bloodmeals suggests either partial feeding and detachment to seek other hosts or transient infestation (phoretic hosts). It is possible that ticks may infest a host and be groomed off before attachment or during the early hours of attachment; either may be sufficient to leave enough host DNA for detection. Then too, ticks will actively detach from dead hosts and can reattach and resume feeding. There has been no means to measure the frequency of these phenomena in nature, and we note that partial feeding could have heretofore unappreciated implications for the enzootic cycle of diverse tick-maintained pathogens.

Our identification of the hosts used by lone star and deer tick nymphs collected from the same site, from the same drag, underscores the hypothesis that apparent host preference interacts with local host availability. Lone star tick nymphs (fed as larvae) exhibited a narrow host range with a clear preference for deer over mice; on Robin’s Island, mice were very rarely identified as hosts for these ticks. Birds can be a common host for lone star ticks (26), and they were commonly found to be bloodmeal hosts for lone star ticks in one site on Prudence Island but not at the other. The Schoolhouse samples were collected from a walking trail edge and those from Blueberry Hill were from the understory of a hardwood canopy; it may be that the abundance and diversity of birds differs between them. Why birds were not found to be hosts for lone star tick larvae on Robin’s Island but were on Prudence Island remains to be explained. It may be that deer are less dense on Prudence Island (the herd has been declining there for the last decade [29]) and, therefore, are less available to serve as hosts for subadults but sufficiently common to maintain the lone star tick population by feeding enough adult ticks. We did not attempt to capture birds and do not know whether lone star tick subadults infest them similarly on the two islands, but different bird species assemblages might also account for the differences if there is a preference for certain birds. Squirrels, shrews, voles, and rabbit bloodmeals were not commonly detected from lone star tick nymphs, although it is possible that these animals were at low densities during the last few years on Prudence or Robin’s Island: deer tick nymphs also did not demonstrate any evidence of having fed on them at these sites either. Previous analyses of host bloodmeal identification in lone star ticks, using a 12S mitochondrial ribosomal DNA (rDNA) reverse line blot method with specific probes, found that squirrels and diverse birds were hosts for these ticks (21, 27), more so than deer.

In contrast to the lone star ticks, the deer tick larvae in this study fed on a greater variety of hosts. There appeared to be a definite hierarchy of preference, with mice being highly favored and deer being least favored. The bloodmeal identification data derived from lone star ticks indicate that deer were readily available on both Robin’s Island and Prudence Island but few Ixodes larvae fed on them. Instead, mice and birds were more likely to be utilized as hosts, suggesting the possibility that deer tick larvae actively avoided deer in order to feed on mice. Accordingly, despite deer being present at these sites, they would not feed enough larval ticks to be zooprophylactic. By contrast, on Nantucket Island, deer were found to have fed more ticks than mice. Assuming that the deer ticks’ preference for mice remains the same regardless of site, it is likely that mice were largely unavailable to serve as hosts on Nantucket during our study; indeed, we found only 9.2% of the (bleached) ticks with evidence of having fed on them. Our limited trapping data demonstrates that mice were scarce in our Nantucket site during the fall of 2020 (when the nymphs collected in 2021 were feeding as larvae): only 4 mice were trapped in 80 trap-nights, compared with 21 in 130 trap-nights on Robin's Island. When their preferred mouse host is scarce, larval deer ticks fed on less-desirable hosts, such as deer. This phenomenon of host-switching and redistribution has been described previously (30) but is not universal, inasmuch as the related Ixodes pacificus appeared to be unable to find suitable hosts when their preferred host, lizards, were removed from their habitat (31).

On Nantucket, deer may be zooprophylactic or dilution hosts, feeding as many as 40% of the deer tick larvae, many more than the other hosts, including shrews. It may be that deer are more likely to reduce the force of transmission on islands with a depauperate fauna, such as Nantucket (11). Zooprophylaxis, however, was not detected: prevalence of B. burgdorferi in host-seeking nymphal deer ticks from Nantucket was not significantly less than that of the other the sites included in this study (Table 6). Furthermore, we found no evidence supporting an association between the proportion of ticks that had fed on deer as larvae and the prevalence of B. burgdorferi infection in those ticks (Fig. 1). It may be that our sampling was underpowered to detect a modest association, but there were sufficient samples to detect a positive association with the proportion of ticks that had fed on mice (Fig. 1). In general, mice are common reservoir hosts for B. burgdorferi, although there are spatiotemporal differences in their contribution. We conclude that deer are only likely to be zooprophylactic hosts in communities with few alternative hosts and that their primary effect would be to increase the force of transmission of tick-borne pathogens by increasing the density of ticks by virtue of their role as the main reproductive host for the deer tick. This hypothesis is consistent with the wide variation in the detected “dilution effect” reported in the literature and emphasizes the importance that local community composition as well as host preference has on the enzootic cycle of tick-borne pathogens.

MATERIALS AND METHODS

Tick collection.

All ticks were collected by dragging using tick drags that had been used for more than a year; the same tick drags were used for all sites, and all sites were sampled by the same person. Tick collections were targeted to deer tick nymphs; lone star nymphs were collected from the same drag samples. Collections were made on Nantucket Island, MA; Robin’s Island, NY; and two sites on Prudence Island, RI, during May and June of 2021. The first two sites were previously described (17). Briefly, our Nantucket Field Station site is within a 43-ha property with dense brush (highbush blueberry, poison ivy, fox grape, greenbrier) and interspersed mowed paths. Robin’s Island is an unpopulated privately-owned 1-mi² island with mature coastal hardwood and a greenbrier and low shrub understory with areas of mown fields. Prudence Island is a 1,443-ha island in Narragansett Bay, RI. Lone star ticks have infested Prudence Island since the late 1980s (32); our Schoolhouse site comprises a nature trail within an oak-pine forest, with a dense greenbrier understory. The Prudence Blueberry Hill site comprises a blueberry shrubland and mixed hardwood forest with a relatively open understory of poison ivy and greenbrier; ticks were dragged from the understory. Ticks from each collection were stored together in a single vial at −20°C in the lab until use.

Tick processing.

Experimental ticks were externally decontaminated using a solution of 5% bleach and then washed 3 times in distilled water before DNA extraction. Control ticks were processed directly with no washing step. Ticks were homogenized individually, and DNA was extracted using 50 μL of QuickExtract for DNA (Lucigen) following the manufacturer’s instructions as follows: 65°C for 6 min then 95°C for 2 min. Deer ticks were tested for B. burgdorferi using a previously published assay (33). Real time PCRs for bloodmeal analysis were done on all ticks as previously described (17, 34). These PCR primers target mammalian retrotransposon DNA remnants from the host that the tick had fed on during the previous stage. All ticks were tested for mouse, deer, bird, squirrel/chipmunk, rabbit, shrew, and vole. Extreme care was taken to avoid contamination of samples. These measures include dedicated clean forceps for tick manipulations, areas for pre-PCR manipulations physically separated from PCR machines, and UV-decontamination of a dedicated PCR dead-air hood between runs. A detailed protocol has been deposited on the Nature Research Protocol Exchange (https://doi.org/10.21203/rs.3.pex-1736/v1).

Bleach validation experiment.

To determine whether 5% bleach could eliminate gross external contamination of ticks, 10 lab-reared nymphal ticks that had fed on Peromyscus as larvae were placed in a tube containing deer hair. Ticks and the hair were mixed thoroughly to ensure every tick had been in contact with the hair. Half were then washed with 5% bleach before extraction and half were not. The ticks were then tested for mouse and deer using our previously described bloodmeal identification assay.

Data analysis.

Differences between identified bloodmeal hosts and bleach treatments were assessed using Fisher’s exact test using the StatPages online calculator (https://statpages.info/ctab2x2.html). Association analysis was conducted within the software package GraphPad Prism. Significance was assigned a priori as P ≤ 0.05.

Data availability.

Data has been deposited in OSF and can be accessed using the following link: https://osf.io/dh5bw/?view_only=5bb2c1d96f464ecfa5608e1f1f6bf820.