192 



\V. i :. AI.LEN. 



growth it is worth while to give them a little general discussion. 

 There are sources of error, which being well recogni/ed, have 

 been checked upon and guarded against as much as possible. 

 (i) The measurements were made and the segments counted 

 without the use of a mechanical stage. Since the specimens 

 possessed very few good guide marks there may have been slight 

 errors due to this cause. (2) Only one count was made in 

 simpler cases although several counts were made in doubtful 

 cases. (3) The correct measurements may be misleading at 

 times because of variability in the speed of regeneration due to 

 lack of uniformity in the bisecting cuts. This is due- largely to 



TABLE II. 



Limnodrilus. 



A copy of the averages from which the growth rate for the first and t-nth 

 regenerated segments was calculated. A partial revision of the averages for th<- 

 total regenerated part is also included here. Measurements are in millimci 

 ,R=length of regenerated tissue. iRS, loRS, etc., first, tenth, etc., regi-nrratol 

 segments. 



Explanation of Table II. Table II. is a copy of the averages, in niilliim i 

 from which the growth rate for the first and tenth regenerated se.nmcnts \vas cal- 

 culated. A partly revised list of the averages for the total re{n-iu-i.iti-i| pan is 

 also included here. A comparison of Tables I. and II., especially cotiMdrimi; thr 

 measurements of old segments in Table I., will give some indication of one disturl 

 factor in the polygon, i. e., varying sizes or extension of the w.um- at tin- mm- o) 

 killing. 7?=total regenerated tissue; i RS, 10 RS, etc., = in-t. until, etc., re- 

 generated segments. 



the smallness and activity of the worms which made it i 

 to cut precisely across or at the same level in the segments. 

 Thus some cuts fell on the inter-segmental line, other- .n various 

 distances within the segment, and several \\rre diagonal. The 



