68 BOTANICAL GAZETTE [JULY 
by De BRUYKER,”' in which it is shown that the means and quartiles, and indeed 
all the ordinates, of the ogive, when calculated according to the methods of GAL- — 
TON, are a half-unit too low, due to the fact that GALTON constructs his curve — 
on the extremity of the classes instead of on their mid-values. The suggested 
correction brings a very close agreement between the median and the arithmeti 
mean. DARBISHIRE?? gives an interesting popular discussion of correlation of the 
kind dealt with by biometers, using as his basis the dice-throwing experiment 
of WELDON, in which the relation between two successive throws of dice are 
the correlated quantities considered, when a certain number of dice from the 
first throw are left on the table to influence the result of the second. When none 
of the dice are left, the two throws are totally unrelated and the correlation is 
zero; and when all are left, so that the second entry is an exact repetition of the 
first, correlation is complete. Tables are given showing the actual results of 
throwing 12 dice in this way under every possible condition-as to the number of — 
dice left back to influence the second throw, thus beautifully illustrating the 
different degrees of correlation between zero and unity — GrorcE H. SHULL. 
Symbiosis.—A recent paper by KEEBLE of Reading and GAMBLE of Man- 
chester, England, continues the investigations of these two naturalists into the 
subject of the symbiosis of plants and animals.?3 Of late years these two men 
have been almost the only ones continuing this study, which, as the literature 
shows, was being actively pursued fifteen or more years ago. 
Convoluta roskoffensis is a simple flat-worm living between the tide-marks 
on the northern coasts of Europe. It is usually green, owing to the presence in 
its tissue of chlorophyll-containing cells, which have been diagnosed as one of 
the Chlamydomonadeae. They resemble the members of the genus Carteria, 
but for certain reason KEEBLE, the botanical collaborator, hesitates to place them 
positively in this genus. The life-history of both components of the association — 
been worked out. From this study it is clear that the flat-worm begins ie 
existence free from green cells. There is no transmission of green cells or evenof 
rudimentary cl topl (plastids) from parent toegg. The green cellsappeat 
to be chemotactically attracted to the eggs and egg-cases. Pure cultures of ge 
alga may often be obtained from egg-cases which have attracted and become — 
partly filled with the green motile cells. By carefully washing in sea-water fre 
from the alga, colorless Convolutas can also be obtained. These will remail 
colorless or will turn green according to subsequent treatment. If kept in S@ 
2t DE BRUYKER, C., Bemerkingen aangaande de Galton’sche curve. Han@ : 
tode Vlaamsch Nat. u. Gen. Cong., S. 1906. pp. 6. figs. 2. . 
22 DARBISHIRE, A. D., Some tables for for illustrating statistical correlate 
Mem. and Proc. Manchester Lit. and Phil. Soc. 5I:no. 16. pp. 21. diagrams 12. 7 ©” 
pl. 28 Je 1907. : 
*3 KEEBLE AND GAMBLE, The origin and nature of the green cells of Conv : 
roskoffensis. Quart. Jour. Micros. Sci. 51:167-219. pls. 13, 14. 1907- 
