294 MR. G. UDNY YULE ON THE ASSOCIATION 



The associations I find for the whole mass and for the five species chosen are 



Number of plants. 



Whole series. ..... 1094 Q = + '66 '025 



Ipomea purpurea .... 146 = + '90 '028 



Petunia violacea .... 154 = + '90 '026 



Reseda lutea . . . . . 64 = + 74 '086 



Reseda odorata 110 = + '49 '103 



Lobelia fulgens 68 = + '29 '153 



Are these differences significant ? Taking successive differences down the table 

 from Petunia violacea onwards I find 



Probable error 

 Difference. of difference. 



Petunia violacea and Reseda lutea . . . '16 '090 



Reseda lutea and Reseda odorata ... '25 '134 



Reseda odorata and Lobelia fulgens. . . '20 '184 



and for the extreme difference 



Petunia violacea and Lobelia fulgens . . *61 '155 



These figures can leave no doubt, I think, that specific differences do exist as 

 regards closeness of association between crossing and vigour of offspring even in 

 species all normally cross fertilised. The difference between Ipomea purpurea or 

 Petunia violacea and Lobelia fulgens is certainly significant, and not only so but each 

 successive difference in the above short table is greater than its probable error. It 

 must be remembered that we are dealing with a different point to that noted by 

 DARWIN ; he is dealing with the amount of the difference between crossed and self 

 fertilised offspring ; we are measuring the approach towards absoluteness of the law 

 that there is a constant difference. The law is much less absolute permits of many 

 more individual exceptions with some species than with others. 



A curious point is the significant difference between the wild and cultivated species 

 of Reseda. The difference may possibly be due to the cultivated character of 

 R. odorata, but certainly need not be, as the two first species on the list in which the 

 association between height and crossing is '9 are both cultivated species foreign 

 to England. Reseda odorata was also erratic in its behaviour as regards self sterility 

 (cf. " Cross and Self Fertilisation," p. 119 and pp. 336-9), some plants being highly self 

 fertile, others quite self sterile. The offspring of highly and slightly self fertile plants 

 were, however, equally vigorous. 



In the table on p. 295 I have entered the sign of the association in the column on the 

 left ; as I have stated, the probable errors are so large that it seems misleading to give 



which only the tallest of each lot was measured, is included. In other cases I have pooled outdoor-grown 

 plants with plants in pots, taking the average height separately as above, and so on. 



