60 TWELFTH REPORT. 



VARIATION I:N LYMNAEA REFLEXA SAY, FROM HURON 



COUNTY.^ 



H. BURRIXGTON BAKER. 



The shells used in this Avork were a set of about five Iiundred speci- 

 mens of Lymmiea reftexa Say. They were obtained during the summer 

 of 1908 by the Michigan Biological Survey, from the protected, sandy 

 beaches of Saginaw Bay around Sand Point, which is the largest pro- 

 jection on the west side of the '"thumb" of Michigan. They occurred 

 there in large numbers, in shallow water, on floating driftwood and on 

 •masses of decaying sedges (mostly t<ciypus amcricumis) . 



The work was begun to determine whether some peculiar, elongated 

 specimens were simply regular variates at one end of a symmetrical 

 curve or belonged to a more or less distinct variety. If the latter was 

 the case, a bimodal curve would, of course, be expected. 



The greatest length and. the greatest width of the shells were meas- 

 ured and the ratio used was the length divided by the width, which 

 in all cases was between two and three. The measuring was done with 

 a Browne and Sharpe micrometer caliper, but it was found that extra 

 precaution had to be taken in measuring the width, due to errors caused 

 by the long axis of the shell not being exactly at right angles to the arms 

 of the caliper. This difficulty was partly obviated by fastening the in- 

 strument to a block and tacking under it a piece of cardboard with 

 a series of lines ruled parallel to the measuring arms and a series at 

 right angles to them. With the assistance of these lines to orient the 

 specimen, it was estimated that the measurements were accurate to about 

 five one-hundredths of a millimeter. 



In collating the data obtained, it was found that the ratios varied 

 with the length and in the same direction ; that is, the older or longer 

 shells had much higher ratios than the smaller ones. This difference 

 was considerably over one-half as much as the gi-eatest difference be- 

 tween any of the ratios, so that a curve obtained by using all of the 

 specimens would have been so obscured as to be entirely incorrect. 

 These differences have been plotted in Figure 1. 



The curve obtained by using shells of a single length was a compara- 

 tively regular, and quite symmetrical one, wherever the specimens were 

 numerous enough to insure comparative accuracy. This means, of 

 course, that we have apparently a case of "fiutleterminate, chance'' varia- 

 tion, except that the length of the shells represent only approximately 

 the actual time age of the specimens. Probably, however, this error is 

 not verv great, as the shells were all obtained from the same habitat. 

 The variation of those of twenty-two and those of twenty-three milli- 

 meters in length is shown in Figure 2. 



The variation of the ratio with age raises a question in regard to 

 Dr. Adams' work on ''Variation in lo,"- as in that he mentions that he 



iFrom the University of Michigan Museum. 



2Chas. C. Adams, Variation in lo, Proc. Amer. Assoc. Adv. Sci., Vol. 49, 1900. 



