EN THE tropics: III. 163 



could be known accurately, which was not the case in the 

 above experiments. 



It is especially to be remarked that such a process as 

 was actually adopted, by which only the seed parent was 

 capable of exact identification, is quite inadequate for the 

 elucidation of anything beyond the very simplest of heredi- 

 tary processes, such as are found for example in the cases of 

 the yellow and white or red and white pairs of characters. The 

 reason of this is that the supposed uniform strain used as a 

 pollen parent may bear invisible but segregable characters, 

 which, on meeting with certain characters in the strain under 

 examination, may produce different results. Such phenomena 

 are now known to occur in several species of plants, but were 

 less clearly understood at the time when the present experi- 

 ments were begun. 



The present observations must therefore serve as a prelimi- 

 nary survey of phenomena, the complete elucidation of which 

 remains as an interesting occasion for further research. 



3. -RED CROSSED WITH WHITE. 



The red colouration has been a source of trouble to many 

 observers owing to its non-appearance as " xenia." In the 

 experiments last described however (70-72) two successive 

 generations were pollinated by a pure strain with exclusively 

 white pericarps. We therefore have presumably plants of 

 the form (red x white) red x white, and the expectation 

 according to Mendel's law is an equality of plants with red 

 and plants with white pericarps. The actual numbers have 

 been already alluded to incidentally : adding them together 

 we set — 



o" 



Red : 28 + 26 + 18 = 72 , 

 AVhite: 33 + 24 + 1 



18 = 72 ) 

 16 = 73 S 



No case was seen of red among the offspring of plants which 

 were expected to be pure white as regards pericarp, so that 

 the proof that Mendel's law holds good for this pair of 

 characters is established with considerable certainty. 



