348 SEMI-CENTENNIAL OF TORREY BOTANICAL CLUB 
bracts, rays, and disk-flowers for 1900 than for 1903. He attributes 
this to more favorable climatic conditions in 1900 and points out the 
importance of taking differences due to climate into account in 
determining ‘place constants." : 
Clark (то), working with timothy, calculated coefficients of 
correlation for the same character in different years. Following 
Tower he calls these coefficients of ‘ place variation." He corre- 
lates height in 1905 with that in 1906, height in 1906 with that in 
1907, and height in 1905 with that in 1907. Не does the same 
with weight. He finds considerable correlation, which means 
that in spite of varying conditions, such as climate, plants high in 
values one year will be high the next. 
Harris (715) has recently called attention to the different 
values of interannual correlation for different characters. This 
study gives a comparison of data of the same sort for successive 
years by means of correlation tables, as Clark (10) has done іп 
the paper just mentioned. The degree of relation is then expressed 
by the coefficient of correlation. This method has been used 
especially in studies on growth in man (Boas and Wissler 05), 
and Harris refers to work of Pearl and Surface on egg-production 
and Gavin on milk records, where interannual coefficients of cor- 
relation were used. Harris published data for interannual cor- 
relations for fruits of Staphylea trifolia and Hibiscus syriacus in 
which he gets various degrees of correlation according to the parts 
of the fruit studied; for example, for 23 fruits of Hibiscus he finds 
the correlation for 1907 and 1908 to be for sepals and sepals 
+0.46, for bracts and bracts +0.84, for ovules and ovules +0.94, 
for seeds and seeds +0.63, etc. While these studies indicate that 
the degree of correlation may be different for different characters, 
they are not especially concerned with the analysis of changes in 
values for a single character due to such a factor as age. 
VII. EVIDENCE THAT POSITION IS A FACTOR IN PARTIAL 
VARIABILITY 
With the recognition of the existence of partial variability 
there developed further refinement of study which aimed to de- 
termine the relation of position on a plant to difference in number 
of parts. Burkill ('95) gives indiscriminate data for 102 flowers 
