232 MEMOIRS OF THE NEW YORK BOTANICAL GARDEN 



In three instances there were one o r more axillary shoots from the 

 cotyledons. 



In all plantlets the primordial leaves were highly abnormal in 

 insertion, form, or number. 



In 19 1 3 and 1914 numerous plants were available for compara- 

 tive study, and extensive descriptive notes were drawn up. Since, 

 taken as a whole, these showed the same general characteristics 

 as the 191 5 series, which is described quantitatively below, it 

 seems unnecessary to present details concerning them. 



As noted above, there were in the 19 14 culture seven plants de- 

 scribed as morphologically essentially normal. Three of these had 

 purple hypocotyls, indicating accidental hybridization with some 

 variety with pigmented seed-coats. Four of these plants died, 

 three produced offspring. Two of these three were purple-stemmed 

 seedlings which produced dark-coated seeds. These were clearly 

 hybrids due to vicinism. The third plant, with a green hypocotyl, 

 produced white seeds which gave only normal plants in 191 5. 

 Possibly this was due to the accidental introduction of a normal 

 seed. The line is being continued. 



In 191 5 there were 2 plants among the 1,202 studied in detail 

 which had the normal number of both cotyledons and primordial 

 leaves. A few other plants of the same kind occurred among the 

 6,400 plants which were not described so minutely. These 

 numerically normal plants generally have other features of leaf 

 form or texture which indicate that they belong to the race. 

 Some of these individuals will be bred further. In the meantime, 

 I consider them as merely the extremes of a highly variable series. 

 Their presence seems to me at present to qualify in no degree the 

 conclusion that the race is fully constant. 



The first question concerning this tetracotyledonous race that 

 the reader would like to have decided is the nature of its origin. 



While it was secured at the very beginning of a selection experi- 

 ment, it can not possibly be regarded as produced or built uj) by 

 selection. It was merely isolated. It was isolated simultaneously 

 in seven, and probabK' in nine,' indixiduals. The records show 

 that all of these nine plants bejong to a single line, the produce ol 

 plant 139 of 1907. 



' The four offspring seedlings of plants 5,782 and 4,434 dictl without producing seed. 

 Their progeny could not, therefore, be adequately tested. 



