In this MS, the authors review the literature of non-consumptive effects and trait mediated indirect effects of anthropogenic disturbances on wildlife. They find mixed evidence across taxa and that these costs are highly dependent on other factors (e.g., taxa, disturbance type), but provide a theoretical model and framework for understanding these effects informed by predator-prey ecology. Overall I found the review to be well written and informative. The results of the review are less than satisfying as they are mixed and likely depend on a variety of other factors not examined. I have a couple larger suggestions and some minor comments for the manuscript:

General: Many studies described here found no NCE's or TMIE's or results were mixed. It strikes me that many species living among humans or impacted by anthropogenic change might (a) already be fully habituated to those effects so no phenotypic changes is seen, or (b) the studies are only grabbing/studying species that CAN live with the anthropogenic changes and thrive despite them, not those that were forced to leave the area. A species that has suffered to great of NCE's might be gone completely or so rare that no studies could be carried out with sufficient sample sizes. These types of biases in the dataset and review need to be acknowledged and explored more fully.

The authors conducted a full review but do not provide any tables or evidence of the studies in the review. A full appendix of the studies included and what type of effect they found should be provided here.

Examples given throughout need to be checked that they provide details on the type of effect found. For example, P13, L37 describes "reproductive outcomes" but not what kind - positive? negative? This is akin to saying an effect is "significant" without telling us what that effect was. There were several of these in the paper and the full MS should be checked to be sure any example is sufficiently explained.

P7: it might be worthwhile to explore the deception signaling literature here as there are cost-benefit analyses of when to respond to a signal when it is accurate (truly a lethal intention) vs false (non-lethal intention).

P8 L10: Suggest change to "Anthropogenic environmental properties"

P8: Many ideas are discussed in these paragraphs but no examples are presented. It would make the ideas easier to understand quickly if a few simple (e.g., xxxx) were included similar to those on P15 L37-56.

P10 L39: List some examples of keywords.

P11 L3: Clarify what you mean by "tested an effect beyond a phenotypic response"

P14 L43-44: How has fleeing contributed to NCE's here? Please explain a bit more - this is to general of a statement.

Figure 3: Do the widths of the curved arrow paths indicate proportional numbers of studies? That needs to be a bit clearer in the legend and figure. Maybe include some summary numbers?

Box 1: The figures presented here are VERY difficult to interpret given the written explanations. I'm still not sure I grasp them fully as the lines are not labeled on the graph and it's not clear why some are even present. I recommend that the box go into much greater detail on each "a" panel in the figure and explain each element as labeled. Why do the blue bars start from some imaginary level at the top of the graph? Where is the "optimal" level of response for each example and why is that the optimal one? Why does perceived mortality decrease with phenotypic change? It seems like if they perceive greater mortality risk, phenotypic change should increase. It also feels like this box should come earlier in the manuscript as it is the framework for the whole study.

The authors have done a fine job addressing most of my earlier comments. Where they haven't, I address them specifically below. Box 1 and the figures is much more interpretable now, but I still have some questions and concerns about the figures. Overall though the paper is understandable and a strong contribution to the literature.

L164-6: I appreciate this addition of the signal detection theory here, but this isn't not much of a mention of it. I was hoping that the authors would use the SDT ideas/literature to support their use of similar perception error effects here. I suggest fleshing it out a bit more in a few sentences.
L373-5: Thank you for this clarification, but fleeing could increase risk of injury during flight or expose the animal to greater risk/exposure to actual predations during flight. So there are some direct costs of flight beyond energetic or foraging effects.
Figure 1: I understand this figure but not sure how much I get from it. I think you could probably cut everything below the NCE and TMIE level, but I'll leave it up to the authors
Figure 2: Indicate direction of effects on each. For example, on the left middle panel, put a + above the first arrow and a - above the second arrow.
Figure 3: Increase font sizes
Figure 5a,c: Why are the NCE bars extending ABOVE the maximum growth rate? Shouldn't the NCE be the reduction in growth rate from the maximum value (x=0 or no phenotypic change)? It feels like there should be a horizontal dashed line starting where the growth curve hits the y-axis (growth curve maximum) and extending horizontally across the figure. And the NCE bars should start there and extend down to the line.
Figure 5b: Why don't the hashed bars on the left match the relative sizes of the hashed bars from 5a? Should the difference in cost between perceived and actual optimal responses be much larger, as represented in 5a?