194 W. E. ALLEN. 



last point seems to be especially important as Miiller (1908) and 

 others have shown that diagonal cuts produce marked differences 

 in the detail of regeneration. 14) The state of contraction at 

 time of killing was also probably somewhat variable so that 

 detailed comparisons would have to be made \\ith caution. The 

 effort was made to test this factor b\ measuring tin- regenerated 

 parts with the old segment as a unit. This indicates greatest 

 speed of regeneration on the tenth day instead of the ninth 

 as indicated by the other method. (5) The small number of 

 specimens available gives opportunity for undue prominence in 

 the showing of individuals. Evident extremes \\ere omitted 

 from the averages. But, after all, the general effect of most 

 of these factors would be to diffuse rather than to accentuate 

 the polygon, and there seems to be ample warrant for saying 

 that the greatest speed of regeneration is at the ninth or tenth 

 day, which is near enough for present purposes. In close ap- 

 proximation to these results we have those from some unpul dished 

 work by Mr. Frank L. Pinckney, using Limnodrilus in this labor- 

 tory in the spring of 1910. As he used the method of successive 

 measurements of living animals, the similarity seems to be 

 practically conclusive. 



In examining the polygons two striking points appear which 

 are rather difficult to explain. In the first place, there is an 

 apparent slackening in growth almost to the zero point on the 

 tenth or eleventh day, the most obvious explanation for which 

 seems to be in an accidental assemblage of extreme characters. 

 At any rate there was no change in temperature or other observ- 

 able laboratory conditions at that time that could possibly ac- 

 count for it. In the second place there is a rise in tin- later 

 portion of the polygon. This, however, in the total regeneration 

 seems to be due to the fact that there are still some 50 or (>o 

 very young segments, main- of which are just at their maximum, 

 thus serving to largely balance the decline of the < ilder regenerated 

 segments, which is not in itself, so very great by that lime. As 

 the study of tissue differentiation has been made principally on 

 regenerated segment no. 10, the polygons lOi segments no. i 

 and no. 10 are shown (Chart I.). The inaMiuum for no. 10 

 appears here at the ninth da\ Inn is apparently only more or 



