Safe Passage or Hunting Ground? A Test of the Prey-Trap Hypothesis at Wildlife Crossing Structures on NH 44, Pench Tiger Reserve, Maharashtra, India
Round 1
Reviewer 1 Report
This interesting paper sets out to test the prey-trap hypothesis at nine crossing structures under national highway 44 in Pench Tiger Reserve, Maharastra, India. It does this by examining the latencies of predator-prey, prey-predator, and prey-feral dog pairs and comparing them to control data within the Tiger Reserve. I consider that the manuscript will be an important contribution to the literature but first ask that some issues be given further consideration and explanation in the manuscript. Also, I would like to see an image of one or more of the crossing structures (during the daytime) included in the manuscript. This will be helpful for the reader to understand the study area and the specific design of the structures. I do understand that the crossings vary in design so a 'typical' one could be used.
An overarching question that I have is whether camera trapping is the best method for testing the prey-trap hypothesis? It assumes that the camera trapping array captures all the important potential interactions at a crossing structure. While camera trapping is a convenient approach logistically, are there better options that could be used to assess this topic? I’d like to see this question being addressed in the Methods and/or Discussion.
The three hypotheses (a, b, c) as currently constituted come across as being somewhat naïve. What do we expect from the literature in terms of the latencies when predators are hunting prey? How is it possible to determine whether crossing structures are acting as a prey trap? Some thought is needed on this subject.
Introduction, lines 44-47, prey-trap hypothesis. Here you describe the hypothesis but then go on to state that there is little empirical evidence that predators focus attention around crossing structures. My question then is where does the hypothesis come from? What is the basis for it/why was it first proposed?
Materials and methods, Study area, lines 73-75. It will be useful to know whether there are only nine available crossing structures or whether there are more crossing structures along highway 44 and you chose a subset of those available for the study. If more than nine was available, please describe how you chose the ones that were used in the study.
Materials and methods, Field methods, lines 89-90. The control sites need to be more thoroughly described. In particular, the location of control sites in relation to the crossing structures should be described (including distances). The spatial independence of sites is important to establish. It is also important to know if the camera trapping approach within the Tiger Reserve produced comparable data to that used at the crossing structures.
Discussion, lines 227-228. Is it wise to increase vegetation cover near crossing structures? My concern is that it provides opportunities for predators that hunt by ambushing prey. Some thought will need to go into the design of revegetation plans.
Minor change to consider on lines 30 and 31: perhaps change “plying” to “driving”.
Author Response
Please see the attachment.
Author Response File: 
Author Response.pdf
Reviewer 2 Report
Manuscript entitled ‘Safe passage or hunting ground? A test of the prey-trap hypothesis under wildlife crossing structures on NH 44, Pench Tiger Reserve, Maharashtra, India’ tested prey-trap hypothesis. Although I found topic of the paper very interesting, and such results should be published, presented version of manuscript must be improved. I suggest major revision of the manuscript. The methods have to correspond with aims and results have corresponded with both, aims and methods. While authors wrote three aims (based on latencies of three different sequences: prey-predator, predator-prey, prey-dog), methods included geometric means of latencies and generalised linear mixed models examining how environmental and crossing structure-related features influenced the latency between sequences. In results, authors showed latencies and geometric mean latencies, and differences between crossing structures, while it wasn’t explained in any other part of the paper. All comments are listed below:
Keywords shouldn’t be used in the title of the paper. Why authors used name of only one focal species? Prey-trap; mitigation; road effects; wildlife; NH44; tiger; conservation; underpass
- 13-14 and later What do you mean by under crossing structures? Under overpasses or under underpasses? For me it is confusing.
- 30-32 I don’t understand what kind of emission from traffic authors meant; moreover, I suggest changing the whole sentence (for example Traffic causes behavioural aversion and avoidance by wildlife [6], [7], and animal-vehicle collisions, one of the leading causes of animal mortality in human-dominated landscapes [8]. )
- 61 I didn’t find the aims of the research the same as in Ford and Clevenger (2010):
(1) Wildlife crossing structures increase predator foraging efficiency; thus, sites where predators kill prey (kill sites) are closer to the highway after construction of WCS. (2) Predators systematically forage near or within WCS; thus, the density of kill sites at WCS is relatively higher than the density of kill sites in the surrounding landscape. (3) Prey activity at WCS increases the probability of an attack by predators; thus, prey use of WCS may be used to predict the probability of predator passage;
I suggest removing Ford and Clevenger 2020 from line 61.
- 74 There is a big difference in cs width! I’m not sure if large structure which might be also passage for wildlife should be analyzed like normal underpass, which width is under 100 m (however, when I think about underpasses I think rather about width under 50 m).
l.83 I understand that not all crossing structures, because underpasses may have curvature shapes. Authors should provide descriptions for all crossing structures, including location, type, and all parameters.
2.3. Analytical methods What did authors mean by pair of sequences (l. 100), sequences (l. 101) and sequences within pairs (l. 102)? Maybe pair or sequence (following Martinig et al. 2020) would be enough?
How does GLMM correspond with the aims of the manuscript?
- 115-117 Did authors mean: Geometric mean latencies for both prey-predator and predator-prey sequences for underpasses were significantly lower than geometric mean latencies for these sequences in control habitat (Wilcoxon rank sum test p < 2e-16).?
Figure 2 Why prey-dog boxplot is so wide? Y-axis: latency (minutes) is enough – remove 'in' in the bracket. I understand that boxplots showed mean, not median, like they usually do?
- 124 I suggest rather write Geometric mean latencies for prey-predator sequence/pair
- 127 Did authors mean sequences or pairs by groups?
Why values for geometric mean and latency (tables/figures) are different? What is difference between these two values and which one was calculated in the paper? I understood that authors wanted to calculate geometric mean, not both. In that case I suggest removing both Figures (2 and 3). Why in tables not all means are listed? In tables, latency is not shown in minutes.
- 143 Larger or wider?
Figure 4. This figure doesn’t correspond with the text above it. The names of crossing structures don’t say anything, as they weren’t explained/shown on map or listed before. Moreover, I suggest preparing boxplots for all MNB in total and all AUP in total.
- 150 Latencies or geometric mean latencies?
The results should be re-write. This paragraph is hard to follow, as 1. Authors didn’t describe the differences between crossing structures (underpasses), especially width of the structures, which seems to be very important in the paper, 2. Authors gave figures with values not corresponding with described methods, 3. I cannot see the results of GLMMs, what should be done according to methods, and what is supposed to be one of the main results, 4. I don’t agree with Authors that data for MNB and AUP should be pooled. I rather suggest analyzing them separately.
- 160 I don’t understand what ‘natural latencies’ are. I supposed that authors thought about geometric mean latency in control sites. Moreover, latencies or means latency?
- 167-168 Present evidence suggests that species use of crossing structures is dictated by factors other than predator or prey activity (Saxena et al., unpublished data) – the is so much data on this topic, you don’t have to cite only unpublished data.
- 167-171 That is not connected with the topic of this manuscript.
Such changes will be connected with changes in abstract and discussion.
I wish authors all the best!
Author Response
Please see the attachment.
Author Response File: Author Response.pdf
Round 2
Reviewer 2 Report
I don't have any other comments.
