To increase transparency, PeerJ operates a system of 'optional signed reviews and history'. This takes two forms: (1) peer reviewers are encouraged, but not required, to provide their names (if they do so, then their profile page records the articles they have reviewed), and (2) authors are given the option of reproducing their entire peer review history alongside their published article (in which case the complete peer review process is provided, including revisions, rebuttal letters and editor decision letters).
The reviewers and I appreciated the careful revision of your manuscript, which is now Accepted for publication.
I am pleased with the modifications/explanations provided and have noting to add recommending acceptance.
I think the authors have addressed all my comments suitably.
One thing they might still add is regarding the point below: (maybe just state this in the discussion.)
The reviewer states, “p12, line 253-262: while this is surely true, it does not seem to be remarkable that structures which mask an insect from direct contact with contact insecticides reduce mortality dramatically.”
RESPONSE: We agree, but this is not a trivial finding because it suggests that there is little to no “backflow” of insecticide as it moves in the air across the swath width. We know this because we only covered the front of the cage.
The authors corrected my in saying that their reporting of medians and means were correct - they are indeed correct, my apologies for the original comment
The authors appear to have done a good job of revising the manuscript.
Thanks for your resubmission. Three reviewers have now evaluated your appeal manuscript.
Reviewer 1, which evaluated the earlier version of your manuscript, is now satisfied and appreciated the quality of your revisions.
Unfortunately, I was unable to obtain a further review from the reviewer who earlier suggested the rejection of your manuscript. As a result I asked additional reviewers to comment, and those two new reviewers have also inspected your study. Their comments are mostly positive, so I am pleased to reconsider my earlier decision and anticipate that you will be able to accommodate the feedback.
Please revise the manuscript addressing the reviewers' comments.
Again, thanks for submitting to PeerJ.
I hope that you appreciated the timely handling of this second revision round. Feel free to contact me if further information is needed.
I am satisfied with the authors's responses to my early comments and their efforts to make this revision. I have no more comments.
MS title: Determinants of acute mortality of Hippodamia convergens (Coleoptera: Coccinelidae) to ultra-low volume permetrhin used for mosquito management
MS authors: Robert K D Peterson et al.
The contribution explores the potential non-target effects of ULV spraying of insecticide against mosquitoes. The field study was carried out in two years and the effects of distance and application heights, together with surface coverage, were considered using two alternate formulations of permethrin. Height exhibited significant effect with higher mortality at ground level, what was also observed when uncovering the cages.
The revised contribution and explanations provided allowed for an interesting reading. The contribution includes two elements of novelty that draw attention – the formulation testing, particularly ultra-low volume, against a non-target insect of a mosquito targeted-spraying effort. Overall, my views are closer to reviewer 1 and 2 rather than reviewer 3, whose concerns were addressed by the authors in detail in their reply. Regardless, the revised contribution is improved and I have little to add, while encouraging its acceptance. My main concern is the downgrading the relevance of surface contact while overemphasizing direct hit of flying insects.
Lns 42-43: I do not dispute the fact that the sprayed droplets can knock down and kill flying mosquitoes. However, the main goal of ultra-low volume formulations is to provide good coverage with minimal volume and the small droplets will assist allowing better surface coverage (and eventual mosquito contact with treated surface), besides of more efficiently reaching the flying mosquito. Surface contact seems also important for mortality in addition to direct hit by droplet, whose efficacy is still relatively small. Thus I would suggest the authors to consider reworking the text to reflect this. Maybe “…tin very small droples (…) to allow maximum coverage with minimum volume leading to high contact exposure of adult mosquitoes and easier reaching flying mosquitoes”; I do recon that the authors targeted the direct hit of insecticide on the insect, but recognition of insecticide residue-impregnated surface is important to prevent oversimplification of the application effects;
Ln 49-50: Along the lines of previous comments I would suggest a change here to something like “…night and also to target the coverage of foliage and other surfaces that will allow the mosquito contact with the insecticide”
Lns 268-271: both things will take place and I would not neglect the surface-contact effect;
Lns 271-272: the text seems imprecise here, and also the context of the references cited leading to a downplay on the importance of surface coverage; I would suggest rewording to enhance precision (e.g., soil coverage is necessarily low under desertic conditions since few droplets will reach the soil with UVL spraying…)
Lns 313-321: the text again downplay the relevance of surface expore while overemphasizing direct interception of insecticide droplets by insects, whose importance is particularly small in agricultural settings. A series of conditions may explain the indicated differences. A recent review provides a glimpse on the potential effects and interactions involved (e.g. Guedes et al. Annu. Rev. Entomol., 2016), which do go far beyong the direct hit in flying insects. I would suggest revision.
This seems like a small piece of a much bigger picture. Ultimately, would it be possible to develop risk-assessment for non-target effects by including some measure of habitat complexity in a treated area, for example, and thus reduced exposure could be predicted? Is that were this work is headed eventually?
I was confused by the way that means and median values were reported. I am not sure that a median of 100% or 0% is possible - unless all the values are 0 or 100....
A larger number of experimental units per replicate would have been good. Although the results do appear to be fairly consistent with the current design.
The results are statistically sound, I believe. The mechanisms behind the effects observed are not as lucid though.
There is no effort made to indicate how these finding could be used in making risk assessments or any decisions about these types of treatments.
The title of the manuscript seems a bit broad. There are many potential determinants of acute mortality that were not examined.
p4., line 69: “such as height of the non-target individuals above ground”
p4., lines 76-89 should really come before the ‘2011 study’ heading, as that part seems to refer to how the insects were handled for all studies.
p5 – replications consist of three stands, each with two cages, one ground level one at 1.5m above ground? This provides one unit of treatment at each distance from the spray line, per replicate. Wouldn’t it have been better to use multiple cages at each distance, to avoid potential issues with uneven distribution of the spray in the air? It must be quite variable over 100m?
p5., line 102 presumably this means that the various treatments were randomly assigned to times or locations in the field?
p6, line 124; p9, line 176 – use km.hr-1 (or km/hr if you must; not kph, no such unit.) (should be corrected throughout document.)
p7., line141: presumably such objects would be things like leaves, branches?
p11, line 218-220: how can the median be 0% or 100%? That seems simply impossible! Same comment on lines 235-236. Same issue in the abstract….
p12, line 253-262: while this is surely true, it does not seem to be remarkable that structures which mask an insect from direct contact with contact insecticides reduce mortality dramatically.
p13, line 274-278: To suggest that the environmental conditions are interacting with the insecticide droplets in unknown ways to influence mortality does not contribute much to the discussion. The results showed that droplet density and VMD were similar at ground level and at 1.5m above ground. Surely this would suggest that the insects are likely to encounter similar doses of insecticide in either position? How would differences in wind speed, relative humidity and temperature be large enough to cause significant differences in susceptibility at ground level vs. 1.5m?
In figure legends – Fig 5 and Fig 6: again, 0% and 100% medians?
As per your email to us, we will now consider your Appeal on the decision.
Please submit your Appeal Letter, and your marked-up revision of the manuscript and we will initiate the re-evaluation.
Three expert reviewers have now read your manuscript. Although Reviewer 1 and Reviewer 2 are quite positive, Reviewer 3 raised substantial concerns that prevent me to support publication of your research in PeerJ.
I hope that the Reviewers' comments will help you to improve the manuscript, before resubmitting elsewhere.
Aerial application of pesticides for the management of mosquitoes is a big concern for the non-target risks to a diverse of non-target arthropods especially endangered species such as butterflies and beneficial insects such as natural enemies and honeybees. The authors have presented a straightforward paper that investigated the non-target risk to a beneficial insect from the ULV spray of a board-spectrum pesticide for the control of adult mosquitoes. Overall, I found this study was well designed and preformed.
The experimental procedures are clear, replicated, and answer the questions as outlined. The 2015 study was obviously improved following the 2011 study. I would suggest to state the source of tested insect in a separate. Right now I am are not sure if the tested insects used in 2015 were purchased from the same supplier as those used in 2011? If they were used at the same ages etc. because of the different control mortality between 2011 and 2015.
Line 215: was the droplets also measured at 20 m in 2015 (but it did not show in Figure 1)?
Their results showed several determinants (distance, height and barrier) on the mortality of tested insects. Most of the data analyses and results are good and the discussion places the work in context with other published studies and does not over reach from the data present.
My only suggestion is that the introduction is too narrow but the discussion is too wordy. The current introduction covers mostly the mosquito literature. In fact, there are many studies on non-target impact of pesticides used for controlling other pests (e.g. agricultural pests) and most recent studies emphasis not only acute toxicity but also sub-lethal effects on the performance (e.g. development, behavior and reproduction) of impacted insects and the non-target effect at the population level (e.g. Stark and Banks 2003, Annu Rev Entomol. 48:505-519). The authors had a lot of discussions on other studies in the discussion section, but the introduction section does not fully cover other studies, for example, human-health risks from residential exposure to mosquito insecticides, non-target risks to other flying insects such as dragon- flies, butterflies and honeybees, or aquatic insects, or other non-target arthropods such as spiders. The manuscript could be greatly improved if justified and developed with a context of other non-target studies (see works by Stark and Banks 2003 and others as examples) rather than sole reliance on empirical data from mosquito studies. The former approach will make the work more interesting and applicable to a wider audience (especially for this journal) while the former restricts it to one organism group.
You maybe give a definition of ultra-low-volume (ULV) applications, as many readers may not fully understand the differences between ULV and other methods such as LV, HV etc.
The paper is easy to read and chosen words are appropriate and in good english.
The objective is clear, but the introduction part need to be expanded in some points: definition of ULV L35.
L53-55 Please expand case studies and if the impact is in the abundance or richness of non target species and species list (or guilds)
L61-62 Please list species traits, and the reason why you chose the ladybug as the best sentinels for ULV impact
The experimental design is in two years but unbalance and different in the two years of research.
It is not well clear the time range between spray application ( not before 19) and the time of the experiment (at dusk) . Have you took into account this variable? Have you look for temporal autocorrelation between different application in the same night?
L89 Please change "approximately" to range of individuals number.
L150 Please change " approximately" or remove.
The statistical analysis seems good for the data collected and hp tested, even if some comparison between 2011 and 2015 are speculative, because not comparable among them.
Results provide some sentences that maybe need to be move to the discussion part L186-187,L196-199, L209-211.
The main conclusion is not speculative. Their findings are supported by the analysis.
L242. Which could be the other factor affecting mortality of H.convergens: the ground cover, the speed of the wind?
I remarked some experimental pitfalls that do not allow me recommending the publication of this manuscript.
Authors used a very weird experimental setup that is half way between a field, a semi-field and a laboratory experiment.
Field researches on the assessment of non-target impact of insecticides are usually conducted on the long run (to follow the population dynamics), on multiple fields and on multiple non-target species. So, this is not a true field study as stated several times in the manuscript.
Semi-field studies usually include a caged natural scenario. Here predators were enclosed on a very small and crawdad arena for a short time period and without a substrate (contemned or not, i.e. refuge area) and or prey (contemned or not).
Non-target toxicity assessment in the laboratory usually includes different route of exposures (eg feeding predators with contemned preys, residual contact and topcial applications) the assessment of sublethal effects, i.e. physiological and behavioral changes on specimens that survived the toxicant exposure. Indeed, it is a pity that authors did not include this kind of test in the specimens brought to the laboratory soon after the pesticide exposure. Most of the article conclusions are based on the fact that caged predators and exposed at the ground level survived the exposure. This does not exactly mean that the predators were not exposed or that they were safe, indeed it is likely that important sublethal effects occurred in those insects (there are several evidences for pyrethrins and coccinellids). I remained even more surprised when I did not find any discussion on this specific issue. Authors seem to have missed all the very important bibliography on sublethal effects of pesticides on non-target arthropods (actually a key topic of the paper).
For the bioassays authors used field-collected specimens of unknown age, sex and unknown pesticide exposure history. Indeed, the outcomes of pesticide toxicology may significantly differ between sexes and in presence of natural stressors (such as age or starvation as in this case), or, even worse, according to previous pesticide exposure in the canola fields were the specimens were collected. This is a crucial point considered the type of tooxicological test performed, i.e., a very short-term spray and count bioassay.
More specific concerns are reported below
Results and their interpretation
According to the literature (authors provided several examples in the discussion) and to the toxicological results of the present study, it is clear that following ULV insecticide sprays non-target organisms' mortality is negligible at the ground level, but cannot be such at higher levels. Thus, authors tested the ULV insecticide non-target toxicity on insect caged at the ground level vs insects caged at 1.5m above the ground (other factors were also included: distance from the sprayer and insecticide formulation). In 2015, authors quantified the insecticide droplets at the various distances and at the two heights.
The main and meaningful finding of the experiment was that volume median diameter and droplet density did not vary between the two heights, so, very likely insects were equally exposed to ULV insecticide at the two heights. Thus, this result did not agree with the toxicological results, i.e. mortality was much higher at 1.5m (85 vs 45%). This is very strange and suggests that something wrong happened in the experiment. Authors did not discussed at all this ambiguity, but just stated "This suggests that there may be another factor or factors contributing to lower mortality of individuals at ground-level compared with individuals at 1.5 m"(L241-242). This is not acceptable for a scientific paper.
The discussion text in L246-248 is really poor and not meaningful, sentences such as "that mortality to non-target insects may vary considerably based on what is in front of them" should never appear in any academic text.
Non-target species model
Moreover, the choice of the insect species seems arbitrary, because authors did not explain (at all) why this species was chosen and what is the likelihood that specimens of this species get in contact with these ULV insecticide sprays. Is this insect species known to fly at dusk? NO
Is this insect species a natural enemy of mosquito adults? NO
So, why using this non-target species?
I am providing some specific remarks not covered in the previous comments.
Title: is misleading. Reading it I was expecting to read something more comprehensive than just the influence of insect position and insecticide formulation on ULV toxicity toward this non-target species
L61-62: more biological and ecological information on this species is needed. Why this species was chosen? What is its important
L63: Number of replications should be properly reported in the manuscript
L67-68: where the insects feed somehow?
L89: approximately is not acceptable, please be more specific. Any idea about the sex of the exposed insects?
L104: when? Where the insect let drying before dong that?
L105: where the insects feed somehow? This kind of predators likes did cannibalisms in such a crowded arena
L160-162: there is a plethora of toxicological studies in which the mortality of insects is expressed as percentage of survived or dead specimens and in which the normality and homoscedasticity of the data distribution are met (with or or without data transformation) and thus parametric data analyses are carried out? Here authors decided arbitrarily (without carrying out the specific tests) that these assumptions are not met! Why
L241-242: this was the unique worthwhile hypothesis to test in this study and authors really cannot leave the discussion so open?
L247: reword Impinge
L300-304: Meaningless. This is not consistent with the authors' findings. Predators were totally exposed to the droplets (no recovery area in the arena)
and "Surprisingly, there were no significant differences in droplet density and VMD between ground-level and 1.5 m. Therefore, reductions in mortality of individuals at ground-level cannot be explained by lower insecticide droplet densities or changes in VMD". I do not see the point of this discussion
L364 and L 369: the fact that individual insect survival is 1 or 0 does not mean that authors cannot provide the mean percentages and thus Standard error of each column. I read tons of toxicological papers with similar endpoints and all of them present mean and SE.
All text and materials provided via this peer-review history page are made available under a Creative Commons Attribution License, which permits unrestricted use, distribution, and reproduction in any medium, provided the original author and source are credited.