> The 300 hundred bee sample for beekeeper
> use is well covered here with plenty of discussion
> of error rates expected.
(Citing Lee, et al 2010 http://dx.doi.org/10.1603/EC10037 I'll call this the "Lee paper" below)
The only error the Lee paper mentions is a "standard error of prediction".
Perhaps someone would like to take a stab at applying the analysis in the Lee paper to a sample of 300 bees, and a finding of 0 mites, 1 mite, and 3 mites.
This will likely lead to a better understanding of the limits of "rolls" vs the desired treatment threshold in terms of mites per bee on a colony-wide basis.
In my view, beekeepers seem to be making treatment decisions based upon samples that are too small to be up to the task of finding less than 10% infestation rates.
Is a 10% infestation rate OK?, Well, what do more recent papers say?
"When varroa infestation levels exceed 2,000–3,000 mites per colony during the autumn season, most over-wintering bees become susceptible to DWV infection, causing colony collapse during the winter. This is because the new brood produced during this period is insufficient to replace the infected over-wintering bees that die" (Barroso-Arévalo, et al 2019)
https://doi.org/10.3389/fmicb.2019.01331
2,000 infested bees is 4% of 50,000 bees. This is the "doomed colony" level of infestation. The colony has no future in fall. But the Lee paper says "Among the 62 intensively sampled colonies, density averaged... 5.2 (SD= 7.8) mites per 100 adults and pupae" 5.2% mites seems close to that 4% that doomed the hives in the Barroso-Arévalo paper I cite above.
I should not have to say more, but a few more points stick out.
The Lee paper goes on the speculate: "There is probably an absolute lowest threshold, where no beekeeper should need to treat. But, between that low threshold and the 10 - 12% level is a gray area where a beekeeper will need to make a treatment decision..."
The 10% - 12% level? Why are they talking about levels double that of the "doomed colony" level? Because the Lee paper was written without any reference to what level of varroa infestation dooms a hive, and they additionally made no attempt to compare its sampling with any actual count of the total number of varroa in any colony. The Lee paper also made a clearly poor guess about the number of mites per cell, saying "If mites invade brood cells at random, then it is likely there were fewer than 1.14 (foundress) mites per infested pupa in our intensively sampled apiaries, which would further reduce the extent of overestimation by doubling." This guess is plainly incorrect to anyone who looks at brood to check for varroa, as multiple foundress mites per pupa are fairly common even at low infestation levels.
I could go on, but I think the message here is that we need to face up to the inherent limitations of a 300-bee sample as a screen for treatment, or the clear statistical evidence that it is essentially a "doomed colony detector", given what we have learned about the impact of the viruses spread by single-digit varroa infestation rates.
***********************************************
The BEE-L mailing list is powered by L-Soft's renowned
LISTSERV(R) list management software. For more information, go to:
http://www.lsoft.com/LISTSERV-powered.html
|