No. 521] 



NOTES AND LITERATURE 



these six indices were calculated. The first point brought out 

 by the measurements is that the wings of queens are larger than 

 those of drones, and the drones in turn possess larger wings than 

 the workers. There are some differences in the proportions of 

 the wings in the three castes, though these differences are not 

 relatively so great as those in absolute size. There are only 

 slight differences, either in absolute size or proportions, between 

 the wings of the two sides of the body. All these are facts 

 easily enough ascertainable by direct observation. The biomet- 

 rical expression makes more precise and accurate what was 

 already known. In regard to the absolute size of the wing the 

 worker is absolutely and relatively more variable than the drone, 

 and the drone than the queen. In regard to the proportions of 

 the wing (as measured by the indices) this relation is nearly 

 reversed, though the differences are much smaller; the drone 

 is slightly less variable than the worker, and the queen is less 

 variable than either. The degree of relative variability, as meas- 

 ured by the coefficient of variation, is exceedingly small for 

 some of the dimensions of the wasp's wing. The mean coeffi- 

 cients for queens' wings are for absolute dimensions 1.57 (right 

 wing) and 1.54 (left wing) and for indices 1.28 (right wing) 

 and 1.25 (left wing). These are among the smallest variabili- 

 ties yet recorded. 



The highest correlations between wing dimensions were found 

 in the workers, and (with the exception of a single pair of 

 characters) the lowest correlations in the queens, the drones 

 taking an intermediate position. The index correlations were 

 found to be nearly all negligibly small. On the basis of this 

 result the suggestion is made that these wing proportions, as 

 given in the indices, ought to be useful taxonomic characters. 



The point of greatest interest brought out by this first study 

 of wasp variation is the relative variability exhibited in the 

 different castes, as com pa rod with what has been found for bees. 

 There can be no doubt apparently that in bees the drones are 

 more variable than the workers. This result was announced on 

 rather dubious statistical grounds by Casteel and Phillips (7), 

 criticized by Lutz (13), but abundantly confirmed by Kellogg 

 (9) and by Pearson's (16, p. 422) reductions of the data of 

 Casteel and Phillips. In the wasp the drones are distinctly less 

 variable than the workers. 



A second study of variation in wasps already referred to has 

 appeared quite recently. The work was done by Thomson. Bell 

 and Pearson (14) and deals with the comparative variability 



