-65- 



Carolina seeds gave intermediate results. Figure 7, from a pho- 

 tograph, ilhistrates the New York and Alabama samples, ten days 

 after sowing. 



Ithaca. Fig. 7— Table 50. Alabama. 



Three other tests were made, wnth the same result. In one test 

 the sample from New York was represented by seed taken from a 

 crib of soft corn, yet this sample gave earliest results, though less 

 marked than in the other instances.* 



CONCLUSION. 



Northern grown corn appears to germinate more quickly than 

 southern grown corn. 



IX. VARIATIONS IN DUPLICATE TESTS 

 UNDER LIKE CONDITIONS. 



It may be well to briefly call attention to the fact that scarcely 

 any two tests made with seeds from the same sample, under con- 

 ditions apparently identical, are exactly alike in results. It fre- 

 quently happens that these results are so dissimilar as to give us 

 no warrant for expressing an opinion of the value of a sample, 

 from two or three tests. The variation in a certain tomato test 

 recorded in this paper, (Cf Nos. 33 and 34), may be taken as an 

 illustration in point. The following table shows the variations be- 

 tween twenty tests : 



*A similar lesson appears to be taught by the behavior of the seeds of spe- 

 cies of Carex, which were planted this spring. Of some 80 pots of seeds, 

 collected by the writer in Europe last year, 13 show germination at the pres- 

 ent time, and of these, all the most forward, with two exceptions, are north- 

 ern species, collected in Scotland. 



