52 University of California Publications in Agricultural Sciences [Vol. 2 



flowers to drop. Thus my only recourse was to label and hag 

 clusters of developing nuts. About 350 clusters were so treated. 

 It is very likely that many of the flowers were cross-fertilized 

 with pollen from neighboring trees. However, if the pollen of 

 either no. 16 or its near-by neighbors had been the source of the 

 new form, qucrcina seedlings should have been obtained among 

 the progeny of other trees besides no. 16 in the original seed test 

 made in 1912. Since only the one tree produces the new form 

 I have disregarded pollination for the present, studying only the 

 pistillate flowers and testing the seeds. 



Cytological investigation will possibly reveal the true nature 

 of the new form and perhaps explain its origin. Meanwhile the 

 speculations concerning the nature of the supposed walnut-oak 

 hybrids are suggestive in this connection. Through failure of 

 pollination or fertilization parthenogenesis may take place. That 

 polyembryony occurs both in the new form and the old is proved 

 by the discovery of the specimens shovMi in plate 19, figures 7 

 and 8. In each case the two embryos were complete, each caulicle 

 being attached to its own pair of cotyledons. Again, it is possible 

 that at some stage in flower development abnormal mitosis oc- 

 curs. Should this happen in a very early stage in the develop- 

 ment of the flower, sufficient abnormal somatic tissue would be 

 produced to make cytological investigation comparatively easy. 

 It is hoped that by determining the morphological location of the 

 nuts that produce quercina seedlings the cytological study of 

 very young flower clusters will be somewhat simplified. 



The present tendency- to refer to hybridization as the basis 

 of all variation calls for a reference to my previous paper^ where 

 I disciLssed the hypothesis of origin through hybridization either 

 with oak or with any form of Jufjiaiis and allied genera, and 

 showed that such a hypothesis is untenable. The results reported 

 in this paper bear out that ccmelusion. Further, as opposed to 

 the proposition of assuming hybridization as necessary for the 

 occurrence of mutation, we have the recent conclusion of Gates^ 

 "that mutation and hybridization are separate phenomena, and 

 that the cause of some at least of the mutations in (EiiofJirra is 

 independent of the combination of hybrid characters." 



