Can Colony Size of Honeybees (Apis mellifera) Be Used as Predictor for Colony Losses Due to Varroa destructor during Winter?

Round 1

Reviewer 1 Report

Manuscript "Can colony size of honeybees (Apis mellifera) be used as a predictor for colony losses due to Varroa destructor during winter?" describes a topic important for beekeeping.

It is necessary to correct the reference in line 76 Meike and Holst 2015 (Meikle and Holst 2015), and the reference Rademacher and Harz 2006, which is on the list of references, but not in the text.

A manuscript “Can colony size of honeybees (Apis mellifera) be used as predictor for colony losses due to Varroa destructor during winter?” describes the results of one study in a series of research of Varroa destructor effects on colony development and survival.

However, this study raised one important question – early identification of colony loss signs that could be recognized and used by beekeepers to allow them to take precocious measures. Validation of simple and inexpensive colony trait is important for its acceptance and practical application. The study had clear research goals. The design of the experiment was set following the given aim of the study. The results obtained through this research indicated the complexity of the problem and the need for further research. The manuscript is well written and easy to read. 

Author Response

Response to REVIEWER 1 (our response is following the >)

Manuscript "Can colony size of honeybees (Apis mellifera) be used as a predictor for colony losses due to Varroa destructor during winter?" describes a topic important for beekeeping.

It is necessary to correct the reference in line 76 Meike and Holst 2015 (Meikle and Holst 2015), and the reference Rademacher and Harz 2006, which is on the list of references, but not in the text.

> Thanks for these suggestions. We changed the text accordingly.

A manuscript “Can colony size of honeybees (Apis mellifera) be used as predictor for colony losses due to Varroa destructor during winter?” describes the results of one study in a series of research of Varroa destructor effects on colony development and survival.

However, this study raised one important question – early identification of colony loss signs that could be recognized and used by beekeepers to allow them to take precocious measures. Validation of simple and inexpensive colony trait is important for its acceptance and practical application. The study had clear research goals. The design of the experiment was set following the given aim of the study. The results obtained through this research indicated the complexity of the problem and the need for further research. The manuscript is well written and easy to read.

> We added in the Discussion the sentence suggested by the reviewer: “Validation of simple and inexpensive colony trait is important for its acceptance and practical application.”

Reviewer 2 Report

Overall, two interesting premises to test (I observe poor colony buildup once varroa exceeds the 5% infestation rate).  On the other hand, commercial U.S. beekeepers regularly observe that it is often the largest colonies in a yard that collapse from varroa, due to their large amount of brood.  But these are far larger colonies than you worked with.Based upon my experience in grading the strength of thousands of colonies for strength during pollination rentals, I must question the accuracy of the "grading from the top" method, despite the supportive Fig. 1 of your 2018 cited paper.  Top grading can be very misleading, since the cluster may be shallow under the top bars.  For single-story colonies, grading is far more accurate by simply tipping the box up and grading from the bottom, also without the need to remove frames.

I also observe the same as your figure 3b, regarding cluster size and temperature.

For your Fig. 1, I suggest that for standard comparison, that you convert to estimated mite infestation rate.  That would be an easy conversion, since a worker bee typically weighs around 110 mg.

The above said, your small numbers of bees in the samples, as well as taking the bee sample from the periphery of the cluster, will tend to underestimate the mite infestation rates of the colonies.

The main weakness of your paper is the small sizes of the colonies involved.  By your own estimation of 1700 workers per frame, coupled with the fact that the frames on the periphery will not be fully covered, you were dealing very small colonies of only 8500 to less than 17,000 bees, whereas a typical full-strength colony contains 40,000-60,000 bees.

I agree with your predictive thresholds for colony death, based upon thousands of my own mite assessments.

Author Response

REVIEWER 2 (our response follows the >)

Overall, two interesting premises to test (I observe poor colony buildup once varroa exceeds the 5% infestation rate).  On the other hand, commercial U.S. beekeepers regularly observe that it is often the largest colonies in a yard that collapse from varroa, due to their large amount of brood.  But these are far larger colonies than you worked with. Based upon my experience in grading the strength of thousands of colonies for strength during pollination rentals, I must question the accuracy of the "grading from the top" method, despite the supportive Fig. 1 of your 2018 cited paper.  Top grading can be very misleading, since the cluster may be shallow under the top bars.  For single-story colonies, grading is far more accurate by simply tipping the box up and grading from the bottom, also without the need to remove frames.

> We would like to thank the reviewer for this comment. In the 2018 paper, we explain the method used and also the correlation with additional observations. We acknowledge the limitation pointe out by the reviewer, but we are convinced that our method is the best available so far with adequate accuracy.

I also observe the same as your figure 3b, regarding cluster size and temperature.

> Great that the reviewer share experiences.

For your Fig. 1, I suggest that for standard comparison, that you convert to estimated mite infestation rate.  That would be an easy conversion, since a worker bee typically weighs around 110 mg.

> Not sure what the reviewer is pointing at. We described our calculation in the Methods section. These calculations were based on our observations. With different weights of worker bees, the overall pattern would remain the same.

The above said, your small numbers of bees in the samples, as well as taking the bee sample from the periphery of the cluster, will tend to underestimate the mite infestation rates of the colonies.

The main weakness of your paper is the small sizes of the colonies involved.  By your own estimation of 1700 workers per frame, coupled with the fact that the frames on the periphery will not be fully covered, you were dealing very small colonies of only 8500 to less than 17,000 bees, whereas a typical full-strength colony contains 40,000-60,000 bees.

> In our study, we used colonies of maximum 20,000 bees to keep them manageable. For the standards in The Netherlands, this is considered as a healthy colony. We acknowledge that bee keepers in different countries (and probably keeping bees for other purposes than research) may have larger colonies. However, given the mite infestation rate, we think that the results will also be valid and applicable for larger colonies. Larger colonies will have a lower chance of mortality (see also the paper of Genersch et al. 2010 in Apidologie). 

I agree with your predictive thresholds for colony death, based upon thousands of my own mite assessments.

> Thanks!