MINUTE STRUCTURE OF PLANT HYBRIDS. 249 



Gynostemium. — I need only refer here to the position of the bundles and their 

 relation to the stigmatic lobes. In C. Spieerianum the three stigmatic bundles run 

 so that lines joining the corners of the stigmatic lobes would pass through their middle ; 

 in C. insigne the bundles would be considerably outside these lines, while in the hybrid the 

 inner margins of the bundles would be touched, or they would lie just without the bundles. 



The stigmatic lobes of C. Spieerianum are equal in size and deep depressions separate 

 them ; in C. insigne two of the lobes are short and one is long (one-half longer than 

 the other two), and except between the short lobes the depressions are shallow ; in C. 

 Leeanum one lobe is rather larger than the other two, and the depressions are inter- 

 mediate in character between those of the parents. 



Pistil. — The ovary of C. Spieerianum is glabrous ; that of C. insigne is densely 

 covered with a short belt of glandular and pointed hairs, which show great diversity in 

 size. In the hybrid the latter description suffices, except that the hairs are reduced in 

 number by about one half. 



(k) Some General Observations on Hybrids. 



Instead of recording all that has been observed regarding the histology of other 

 hybrids, I shall now select some special features which seem to deserve consideration. 



Stomatic Distribution in Hedychium Sadlerianum and its Parents. — The number of 

 stomata over a given area has^ been found in most eases to be intermediate in a hybrid 

 between that of its parents. But on overhauling the hybrid named above, I was greatly 

 puzzled to account for apparent discrepancies in relative distribution, and at first ascribed 

 these to the choosing of leaves that did not exactly correspond in time of development. 

 After exercising every care, however, in this and other respects, the results were 

 discouraging. It should her& be stated that the upper epidermis of H. Gardnerianum 

 has few stomata, and is glaucous owing to a rather thick wax layer covering the cells ; 

 the upper epidermis of H. eoronarium is bright green, devoid of wax covering, and pro- 

 duces a few rather long hairs. The stomata of the latter are ten to twelve times more 

 abundant than those of the former. The upper epidermis of H. Gardnerianum gave an 

 average over equal areas of two stomata ; that of H. eoronarium, thirty-two ; that of the 

 hybrid, twelve. Over the lower epidermis the numbers, as taken from a large series of 

 observations made over the entire leaf surface of several leaves, were 10 : 20 : 22. That 

 this exceptional result was not to be explained in the same way that we, in a later part of 

 this paper, attempt to explain the tendency which a hybrid often has to sway towards 

 one or other parent seemed evident, particularly since other parts of the hybrid are very 

 exactly between those of the parent. 



It appears possible that we may have here a morphological adaptation in the hybrid 

 for physiological work, or, in the truest sense, a case of physiological selection. We are 

 dealing with two parents, one of which develops a thick wax covering to a thick and 

 somewhat leathery epidermis, the other a thin cuticle and scant hair growth from a com- 



