I{. I'kahl and f. J. J)i;nhak 335 



'l"ci linil thal. lilicrc is praet.ic-;illy no foiTrlatimi bclwccii darkncss of coliuir 

 and size in Aj-ceUa shells is an uucxpected and rallicr iiitcrestint^ icsult,, and is 

 a])parcntly contrary to the opiiiion goncraliy Iicld ami)iit( zuologist.s rej^Mnliiiif llio 

 matter. Tliis rcsult sccni.s tu di'tinitt'ly negative any view whicli would niaintain 

 that the Arcella Kliell grows darker witli its cultural age. Or, to express it in 

 anotlier way, this very low correlation coofficient betwccn coiour and size seenis 

 to sliow that \ve should not expect tho (n + l)th or {n + 2)tli generation of Arcella 

 to be on the average darker in shell coiour than the «th. This statemcnt of 

 course assuines what Observation clearly Warrants, that the shells of the (ft+l)th 

 and (w4-2)th generations will be ou the average larger than the (ith generation 

 shells. It must be repeated, however, that it oan only be affirmed that this lack 

 of correlation between coiour and size holds for tlie generations falling within 

 the series of observations here reported. One of the points which impresses 

 one most strongly in following through the history of an Arcella culture is that 

 the very minute shells which appear in numbers in the very earliest part of 

 the cultural developmcnt are all very light in coiour, falling within our coiour 

 class A. These extremely small individuals are probably developed froni spores 

 and form their shells individually entirely de novo. The existencc of thcse small, 

 universally light colourcd shells early in the course of the developinent of the 

 culture would indicate that our result of no correlation between coiour and size 

 would not hold i'or the hrst few generations fi.irnied in the culture. 



On the further question as to whether tho individual shell changes in coiour 

 during the course of its existence, wo have at present no evideuce to offer. 



3. The rate of change of the average size of shell with the age of the 

 culture. 



As has been stated earlier in the paper the measurements cxtended over 

 a period of approximately three weeks. Between the time wheu the first and 

 the last measurements were taken there was evidently an opportunity for the 

 average size of the individuals in the culture to change. The culture was in 

 a flourishing conditiou during this period, and it scemed a matter of sonie 

 interest to determine how much of an increase in the average size of the 

 individuals occurred. 



The Cards on which the measurements were recorded were numbered con- 

 secutively in the order of measuring, the card bearing the record of the first 

 individual measured being numbered 1, the next 2, and so on up to tho last 

 individual. This consecutive numbcriug furnishes a method of ap})roximating 

 to the desired determination. The cards recording the measurements of tho tirst 

 500 indi\'iduals were arranged in 10 classes of 50 individuals each. The first class 

 contained the first 50 Arcellae measured, the next class the second 50 and so on 

 to the tenth class, which contained the last 50 individuals. Thcn the cards in 

 each class were an-anged into arrays with reference to outer diameter. In this 

 way was formed a doublc-eutry table with position in the order of mcasurement 



