158 



SCIENTIFIC AGRICULTURE. 



January, 1922. 



gated by liim the yellow seed averaged 

 lower than the purple (Exp. Sta. Ree. Vol. 

 31, 1914, p. 330.) 



The findings of Bauman and Holde- 

 fleiss . that the weiglit of the seed from 

 vellow-seeded plants is lower tlian that of 

 the seed from dark-seeded plants can, nrw- 

 ever, not be taken as conclusive eviderce 

 of a universal weight and color correlation, 

 as will be ^hown by an analysis of the 

 relation between weight and color made by 

 the writer. The writer collected, in 1911, 

 seed from about 250 individual plants of 

 red clover. The samples were subjected 

 to an analytical investigation with t^e ob- 

 ject of determining the range of variation 

 as far as color, shape and weight of the 

 seed were concerned, and also with the 

 object of determining whether a true cor- 

 relation between color and weight of the 

 seed really exists. The following extract 

 may be quoted {Matte, 1, p. 533). "From 

 the 250 samples collected last summer, 39 

 samples which were either entirely yellow 

 or almost yellow were picked out at ran- 

 dom. The average weight of 100 seeds of 

 those 39 samples was found to be 166.23 

 mg. Of the darkest colored samples there 

 were also picked out 39 samples at random, 

 100 seeds of which were found to weigh 

 167.23 mg. This means that, the differ- 

 ence between yellow and very dark colored 

 being only 1 mg., there does not seem to 

 exist any correlation between the color and 

 the weig'ht of the seeds. That this really 

 is the fact can be clearly demonstrated if, 

 instead of basing t^ie conclusion on the 

 average weight, we study the weight from 

 each single individual. The following 

 table shows the distribution of the above 

 mentioned 39 samples of yellow and of 

 dark colored seed into the different weight 

 classes. It shows beyond contradiction, 

 that the weight of t^he seed varies indepen- 

 dently of the color, i.e., that there does not 

 exist any correlation between color and 

 weight. ' ' 



Although the number of the samples of 

 seed from individual plants classified in 

 the table is small, there can be but little 

 doubt that the figures really are expressive 

 of the fact that the weig'ht of the seed of 

 individual plants is in no way associated 

 witli the color. 



Kaj'anus (3, pp. 770-72) who approach- 

 ed the correlation question from a some- 

 wjaat different angle arrived at a similar 

 conclusion. Kojanus analyzed the figures 

 bearing on weight and color secured from 

 the progenies of mother plants having a 

 known color and average weight of seed, 

 and found that any fixed correlation be- 

 tween weight and color does not exist. 

 Color of Seeds and Agricultural Value. 



It is commonly supposed, by seedsmen, 

 as well as by farmers, that lig'ht-colored 

 red-clover seed is inferior to dark-colored, 

 the result being that the latter is generally 

 preferred to the former. This preference 

 is most often founded on the belief that 

 yellow or otherwise ligjit-colored seed is 

 more or less immature, while dark-colored 

 seed is looked upon as being fully mature. 

 Consequently dark-colored seed is sup- 

 posed to germinate better and to produce 

 a heavier and mtore vigorous stand of 

 clover than light-colored seed. 



Concerning the germination of differ- 

 ently colored seed a few observations may 

 be mentioned. Thus Preyer (1, p. 19), in 

 germination experiments in the laboratory, 

 found no perceptible difference in germin- 

 ation between yellow and purple seed. 

 Similar results were obtained by Beal (1, 

 p. 340) and Kajanus (3, p. 774). In this 

 connection may also be mentioned inves- 

 tigations carried on by Rostrup witfh seed 

 stored for several years. Rostrvp (1, p. 

 23; 2, p. 66; 3, p. '45) found that yellow 

 and purple seed retain their germination 

 power in about the same degree or, which 

 means the same, lose their vitalitv equallv 

 fast. 



In the above-mentioned cases the gn*- 



Mg. 



YelloAv 

 Dark . 



120 130 140 150 • 160 170 180 190 200 210 220 230 240 

 to to to to to to to to to to to to to 

 129 139 149 159 169 179 189 199 209 219 229 239 249 



12 

 5 



