Updated Survey Results: Correlation of Common Pathogens with CCD.
The number of respondents to the BeeAlert Survey has increased steadily. The data base now includes a total of 507 valid surveys received as of 15 April 2007. The survey includes paper and electronically submitted data provided voluntarily by beekeepers who have responded to requests for information at meetings and through apicultural web sites. Responses that are entered into the data base are screened to ensure that no duplicates are entered, but other than that the data are not modified. The data represent the subjective observations of the individual respondents; no attempt at validation has been made. Furthermore, because of the voluntary nature of data accumulation, the data do not represent a statistically designed sampling effort. We have data only from respondents who are motivated to provide the data requested. Beekeepers who are unwilling or are too busy to participate are not represented in the data that are reported below. There also has been no attempt to apportion sampling by size and scope of operation, or by geographical region.
Despite these limitations on the data, the large number of respondents has provided us the most extensive coverage of the current state of the CCD problem of which we are aware. The additional data gathered is also providing us with a broad base from which to explore whether there are patterns of disease, toxic exposure and management practices that may be linked to the occurrence of CCD.
To begin with we review the makeup of the population who contributed the data. Most of the respondents report managing fewer than 100 colonies (66 percent). The remaining third roughly evenly represent larger scale operations. It is valuable that we now have data provided by more than fifty large beekeeping operations that manage apiaries greater than 10,000 colonies. Still approximately two-thirds of the respondents are small beekeepers having fewer than 100 colonies (Figure 1).
Figure 1. Distribution of size of beekeeping operation among respondents to the BeeAlert survey. 507 total responses have been received as of April 2007.
Geographical distribution of the respondents is nearly as wide as reported incidence of CCD (reference CCD map). Respondents did not all indicate the state or province from which they were reporting. The number who did, however included all geographical regions of the U.S. and five Canadian Provinces (Table 1). It is important to the generality of our summaries that wide area coverage is present; it is also important that no one region dominate the survey data. Note that California and North Carolina, which have the largest number of respondents identified, only account for 13 percent and 10 percent of the locations in the data set respectively.
Table 1. Geographical distribution of respondents who listed site locations with their responses to the BeeAlert survey.
Alabama / 2 / 1.1 / 1.1
Alaska / 1 / .6 / 1.7
British Columbia / 1 / .6 / 2.3
California / 22 / 12.6 / 14.9
Colorado / 10 / 5.7 / 20.6
Conneticut / 2 / 1.1 / 21.7
Florida / 10 / 5.7 / 27.4
Georgia / 12 / 6.9 / 34.3
Idaho / 1 / .6 / 34.9
Illinois / 10 / 5.7 / 40.6
Iowa / 4 / 2.3 / 42.9
Kansas / 1 / .6 / 43.4
Kentucky / 2 / 1.1 / 44.6
Maine / 1 / .6 / 45.1
Maryland / 3 / 1.7 / 46.9
Massachusetts / 2 / 1.1 / 48.0
Michigan / 4 / 2.3 / 50.3
Missouri / 4 / 2.3 / 52.6
Montana / 3 / 1.7 / 54.3
Nevada / 1 / .6 / 54.9
New Brunswick / 1 / .6 / 55.4
New Hampshire / 4 / 2.3 / 57.7
New Jersey / 4 / 2.3 / 60.0
New Mexico / 1 / .6 / 60.6
New York / 7 / 4.0 / 64.6
North Carolina / 18 / 10.3 / 74.9
Nova Scotia / 1 / .6 / 75.4
Ohio / 9 / 5.1 / 80.6
Ontario / 1 / .6 / 81.1
Pennsylvania / 6 / 3.4 / 84.6
Quebec / 1 / .6 / 85.1
South Dakota / 1 / .6 / 85.7
Tennessee / 5 / 2.9 / 88.6
Texas / 2 / 1.1 / 89.7
Vermont / 1 / .6 / 90.3
Virginia / 7 / 4.0 / 94.3
Washington / 1 / .6 / 94.9
West Virginia / 3 / 1.7 / 96.6
Wisconsin / 6 / 3.4 / 100.0
Total / 175 / 100.0
Respondents across all sizes of operation indicate high frequencies of severe bee losses over the past six months. Smaller operations are more likely to have suffered average or lower losses than normal(Figure 2). The extent of severe losses increases in intermediately sized operations, but drops again among the largest beekeepers reporting through our survey. Overall, however, 38 percent of respondents report severe losses this past winter. This is roughly equal to the 37 percent who report average or better losses.
Figure 2. Severity of colony loss reported by beekeepers relative to size of operation.
We asked respondents to indicate whether the general cause for demise of colonies was due to overwinter losses, mites, pesticide exposure, or disappearance (CCD). Broken down again by size of operation, the data indicate a high level of disappearance, equaled only by overwinter losses (Figure 3). Colony disappearance and overwinter loss were greatest among the smallest bee operations. Although disappearance is most often indicated as the cause of loss by intermediate and large scale beekeepers, it is not claimed substantially more often than overwinter kill. Interestingly, pesticides are indicated as the cause of colony failure about four percent of the time regardless of operation size.
Figure 3. General cause for colony collapse relative to operation size as reported by BeeAlert survey respondents.
When we compared the reported cause of colony collapse with severity of loss, we observed that average or better loss rates were due primarily to overwinter losses. As severity of loss increased, overwinter losses remained roughly constant. So too, did mite disease, except for a slight increase in severe losses. The impact of disappearance increased consistently as severity of loss increased. When severe loss was reported, disappearance was implicated nearly twice as often as any other factor.
Figure 4. Cause of colony collapse relative to severity of loss as reported by respondents to BeeAlert survey.
The scale and impact of CCD is quite evident in the responses we have observed to our survey. Nearly forty percent of respondents report severe losses. In their estimates of total colony loss as a fraction the total, losses in the severely affected group exceed 75 percent on average. By nearly two to one ratio, implicated cause for these losses is CCD.
One of the postulated causes for the emergence of CCD that has been put forward is an interaction among currently prevalent pathogenic agents. Under this hypothesis, the cumulative effects of a combination of otherwise less lethal agents might be responsible for the phenomenon. We asked for more details about disease conditions that were observed in failing colonies. The pathogens for which we surveyed included:
Varroa mites, tracheal mites, American foul brood, European foul brood, Nosema, chalkbrood, purplebrood, stonebrood, sacbrood, DWV, APV, hive beetle, and wax moth.
Respondents were asked to record the intensity of infestation of each agent on a four-point scale. We performed Pearson rank-correlation among each of these agents and with the occurrence of CCD (Table 2). The resulting correlations showed a number of significant co-occurrences among different pathogens. Overall, relatively low correlations were observed, generally below r = 0.50. That is mostly due to the very large sample size of 495 valid respondents. Consequently, even a low correlation of r = 0.2 is highly significant. The most frequent pattern of significant correlation was a positive co-occurrence of related pathogens. For example, Varroa and tracheal mites appear to occur together as do related bacterial and viral agents. Even given adjustment for multiple pairwise comparisons, the level of significant co-occurrence of pathogens is striking.
For all that, even more striking is the complete absence of any correlation between CCD and any pathogenic agent. This appears to be strong evidence that CCD is not linked to any of the agents listed in our survey, either singly or in combination.
Table 2. Spearman rank-correlations among different pathogenic agents and CCD observed in failing colonies.
Disappear(CCD) / Varroa Mite / Tracheal Mite / American Foul Brood / European Foul Brood / Nosema / Chalkbrood / Purplebrood / Stonebrood / Sacbrood / DWV / APV / Hive Beetle / Wax Moth
Disappear
(CCD) / Correlation Coefficient / 1.000 / .032 / .044 / .088 / .085 / -.017 / .036 / .065 / .071 / .027 / .082 / .063 / .031 / .006
Sig. (2-tailed) / . / .478 / .332 / .051 / .060 / .710 / .424 / .148 / .117 / .542 / .068 / .160 / .497 / .900
Varroa Mite / Correlation Coefficient / .032 / 1.000 / .328(**) / .234(**) / .186(**) / .302(**) / .363(**) / .106(*) / .170(**) / .153(**) / .446(**) / .197(**) / .154(**) / .226(**)
Sig. (2-tailed) / .478 / . / .000 / .000 / .000 / .000 / .000 / .019 / .000 / .001 / .000 / .000 / .001 / .000
Tracheal Mite / Correlation Coefficient / .044 / .328(**) / 1.000 / .273(**) / .219(**) / .303(**) / .264(**) / .171(**) / .115(*) / .234(**) / .258(**) / .166(**) / .056 / .194(**)
Sig. (2-tailed) / .332 / .000 / . / .000 / .000 / .000 / .000 / .000 / .010 / .000 / .000 / .000 / .214 / .000
American Foul Brood / Correlation Coefficient / .088 / .234(**) / .273(**) / 1.000 / .517(**) / .197(**) / .446(**) / .201(**) / .279(**) / .312(**) / .234(**) / .251(**) / .072 / .166(**)
Sig. (2-tailed) / .051 / .000 / .000 / . / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .110 / .000
European Foul Brood / Correlation Coefficient / .085 / .186(**) / .219(**) / .517(**) / 1.000 / .233(**) / .467(**) / .410(**) / .448(**) / .448(**) / .248(**) / .259(**) / .101(*) / .189(**)
Sig. (2-tailed) / .060 / .000 / .000 / .000 / . / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .024 / .000
Nosema / Correlation Coefficient / -.017 / .302(**) / .303(**) / .197(**) / .233(**) / 1.000 / .349(**) / .188(**) / .193(**) / .315(**) / .337(**) / .187(**) / .044 / .181(**)
Sig. (2-tailed) / .710 / .000 / .000 / .000 / .000 / . / .000 / .000 / .000 / .000 / .000 / .000 / .329 / .000
Chalkbrood / Correlation Coefficient / .036 / .363(**) / .264(**) / .446(**) / .467(**) / .349(**) / 1.000 / .227(**) / .327(**) / .379(**) / .352(**) / .196(**) / .021 / .255(**)
Sig. (2-tailed) / .424 / .000 / .000 / .000 / .000 / .000 / . / .000 / .000 / .000 / .000 / .000 / .647 / .000
Purplebrood / Correlation Coefficient / .065 / .106(*) / .171(**) / .201(**) / .410(**) / .188(**) / .227(**) / 1.000 / .692(**) / .503(**) / .234(**) / .407(**) / .167(**) / .149(**)
Sig. (2-tailed) / .148 / .019 / .000 / .000 / .000 / .000 / .000 / . / .000 / .000 / .000 / .000 / .000 / .001
Stonebrood / Correlation Coefficient / .071 / .170(**) / .115(*) / .279(**) / .448(**) / .193(**) / .327(**) / .692(**) / 1.000 / .564(**) / .262(**) / .333(**) / .154(**) / .131(**)
Sig. (2-tailed) / .117 / .000 / .010 / .000 / .000 / .000 / .000 / .000 / . / .000 / .000 / .000 / .001 / .003
Sacbrood / Correlation Coefficient / .027 / .153(**) / .234(**) / .312(**) / .448(**) / .315(**) / .379(**) / .503(**) / .564(**) / 1.000 / .246(**) / .297(**) / .153(**) / .199(**)
Sig. (2-tailed) / .542 / .001 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / . / .000 / .000 / .001 / .000
DWV / Correlation Coefficient / .082 / .446(**) / .258(**) / .234(**) / .248(**) / .337(**) / .352(**) / .234(**) / .262(**) / .246(**) / 1.000 / .359(**) / .195(**) / .107(*)
Sig. (2-tailed) / .068 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / . / .000 / .000 / .017
APV / Correlation Coefficient / .063 / .197(**) / .166(**) / .251(**) / .259(**) / .187(**) / .196(**) / .407(**) / .333(**) / .297(**) / .359(**) / 1.000 / .205(**) / .103(*)
Sig. (2-tailed) / .160 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / .000 / . / .000 / .022
Hive Beetle / Correlation Coefficient / .031 / .154(**) / .056 / .072 / .101(*) / .044 / .021 / .167(**) / .154(**) / .153(**) / .195(**) / .205(**) / 1.000 / .234(**)
Sig. (2-tailed) / .497 / .001 / .214 / .110 / .024 / .329 / .647 / .000 / .001 / .001 / .000 / .000 / . / .000
Wax Moth / Correlation Coefficient / .006 / .226(**) / .194(**) / .166(**) / .189(**) / .181(**) / .255(**) / .149(**) / .131(**) / .199(**) / .107(*) / .103(*) / .234(**) / 1.000
Sig. (2-tailed) / .900 / .000 / .000 / .000 / .000 / .000 / .000 / .001 / .003 / .000 / .017 / .022 / .000 / .
** Correlation is significant at the 0.01 level (2-tailed).
* Correlation is significant at the 0.05 level (2-tailed).