STUDIES OF VARIATION IN INSECTS 331 



aft structural correlations. Do the variations, continuous and 

 discontinuous, show similar bilateral and metameric correlation ? 

 Evidence regarding this question will be found on many pages 

 in the present paper, right and left correlation, at least, having 

 been considered and briefly discussed in connection with almost 

 all of the various cases studied. And the evidence is curiously 

 conflicting. For example in the male black ant in which was 

 studied the variations of the venation and number of hooks, a 

 close correlation in the variation conditons of right and left 

 wings exists. On the other hand in the honey-bee the bilateral 

 correlation of variation seems surprisingly small (see pp. 214- 

 222.) In the case of variations in pattern also there is no uniform- 

 ity among the various cases studied. In Hippodamta convergens 

 (p. 257 et seq.) the two elytra show pattern-variations quite 

 independently ; in Diabrotica soror (p. 274 et seq.) on the con- 

 trary there seems to be a marked right and left correlation 

 in the elytral pattern-variation. In the cases of the variation in 

 number of tibial spines on the right and left hind tibiae of 

 locusts (p. 301) and cicadas (p. 306) we have simply made a 

 brief statement, in each case, of the actual conditions of corre- 

 lation leaving the reader to draw his own conclusions. In the 

 case of the variation in actual and relative length of the anten- 

 nal segments of the scale insect, Ceroputo yucca (?) (p. 310) 

 there is a surprising lack of correlation between the right and 

 left antennas. 



We have not attempted to determine the mathematical ex- 

 pression (coefficient of correlation) for any of the cases studied. 

 The data presented, however, will enable any biometrician who 

 sees an advantage in doing this, to do it. But without check- 

 ing our results by the use of that method there seems, on the 

 whole, to be a surprising lack of that fine degree of correlation 

 in variation which we should expect to find existing if we 

 believe that the actual existing conditions of structure and 

 pattern in these bilaterally symmetrical animals are an expres- 

 sion of the result of the action of a rigorous natural selection. 

 If one condition of pattern or structure is the most advantageous 

 (of the many conditions which selection among a host of fluc- 

 tuating variations could have established) surely this condition 



