2ÖS Growth qf Cell in Daithiiia Magua 



the positions of tlie 3rtl ;m<l 4th ccdyses. In these daphiiia the body length 

 at the Ist ecdysis was .about 36 units, at the 2nd about -AS, at the 3rd 57 — 61, 

 at tlie 4tli 77 — 91 ; after the 4th shcddiii«,' tlie aiiiinal gcnerally prodiices eggs in 

 the brood ponch and there is an indefinite number of ecdyses varying according to 

 the length of life. 



It might porhaps be objected that the nieasurements of the polygonal areas of 

 the carapace simply indicatcs the ain<nnit of stretching of the cxoskeletiin betwcen 

 successive nioults and that the cells divide underneath quite iudependently of 

 them. But if this were so the number of nuclei under each area would gradually 

 increase as the nninial approached its next moult. An examination of a number 

 of individuals did not, however, support this view and it appears that not until 

 the creature is on the point of mouiting do the cells divide to any extent, and 

 then it is clear froin a glance at the diagram that the divided cells are larger 

 than the cells which occurred immcdiately after the previous ecdysis, i.e. the 

 mean of the cells in array 87 — 91 is higher than that of array 57 — Gl. 



In the diagrani the liues of correlatioa are draun thnnigh three successive 

 periods of growth (1) the period from the time of hatching to the tinie whcn the 

 3rd moult is about to take place, (2) the period from the 3rd moult to the 

 4th moult, (3) the period after the 4th moult has occurred. 



Since the individuals are grouped simply according to the size of the body 

 these periods will con.siderably overlap, for the size of the animal at which the 

 different ecdyses take place varies very considerably from animal to animal. 

 This overlapping of the different periods of growth produces a disturbing element 

 which is especially noticeable in the array 82 — 86 and it obliterates the destruction 

 which otherwise should be obscrved between the periods of the Ist and 2nd 

 ecdyses. 



If we make allowance for this disturbing element it niust be agreed that the 

 means of the arrays Cluster fairly weil amuiid the tiiree liues of correlation ; the 

 constants are * 



(1) Correlation r= -öSOö ; Regression = 11337 ; 



Regression line : y - 3 4200 = 11337 {x - 7-6500)t. 



(2) CoiTelation ?•= -3930; Regression = •5851; 



Regi-ession line : y - 8-4329 = -öSöl {x - 9-9217). 



(3) Correlation r= 2477; Regression = -4813; 



Regression line : y — 13-3704 = '4813 {x - 9-5944). 



It will be Seen that the regression is greatest at the first period of growth and 

 least at the third period, that is the size of the constituent cells is less closely 

 related to the size of the body in the older than in the younger periods of growth. 



* I am greatly indebtcd to Dr A. Lee for most kindly calculating these constants ; and also those of 

 tlie parabola to bu mcutioned below. 



t y 18 the nicau celllength, x the body-length, both measured in the working ßcale units of the 

 correlutiuu table on p. 2.37. 



