HYBRIDS AND GARDEN VARIETIES. 223 



generation of hybrids is coming on. All of the hardy yellow hybrids are 

 entirely sterile, and crosses of N. tetragona with Eucastalia types are also 

 sterile ; as examples of these N. laydckcri rosea and N. pygmaca alba 

 Marliac may be mentioned. Even N. alba X Candida is greatly enfeebled 

 in seed-producing capacity. But N. alba and N. a. rubra, though much 

 more distinct than the last two to the casual observer, are perfectly fertile 

 together and have fertile progeny. N. odorata and tuberosa hybrids are 

 not self-fertile, but can be crossed with pure species. In the Lotos group, 

 some forms are sterile, others fertile. Planchon found N. ortgiesiano- 

 rubra to be impotent as to the ovules, but capable of fertilizing N. lotus or 

 N. rubra. In those now in cultivation, N. devoniensis is nearly sterile, but 

 N. omarana, which closely resembles N. ortgiesiano-rubra, often produces 

 seed. In general, the nearer the hybrid approaches in color to the highly 

 fertile N. lotus, the more productive it is. Several seedlings of N. oma- 

 rana have been raised, of which N. Geo. Huster deserves special mention. 

 Its flowers are of an extremely deep dark red, more so than N. rubra 

 itself, as cultivated here. If the original seed was fertilized by pollen of the 

 same flower, which it is impossible to know, we would have a case of 

 "separation of characters" according to Mendel's theory; but even if 

 N. rubra was the pollen parent, the determining elements of the color of 

 the flower in this seedling seem to have lost all trace of the influence of a 

 white ancestor. 



We have expressed a belief in the occurrence of variations under 

 cultivation without mixture of blood. Probably N. odorata exquisita is such 

 a variant from N. odorata rosea, as also N. froebelii from N. alba rubra, 

 N. devoniensis from N. rubra, and N. gladstoniana from N. alba ; A 7 , glad- 

 stoniana is quite fertile, and has seeds and arils like N. alba. Whether 

 N. alba candidissima belongs in this class or among the hybrids is uncer- 

 tain. It is absolutely sterile, and may be a cross of N. Candida with the 

 large N. alba of Greece. Quite comparable with these varietal changes 

 in natural forms are some suddenly appearing peculiarities of hybrids. 

 For example, N. robinsoni can be easily recognized by a peculiar crimped 

 notch and fold about the middle of each side of the sinus of the leaf; 

 N. gloriosa is uniformly pentamerous ; and one of a set of plants of 

 N. gladstonianaKmexicana raised by Mr. Tricker habitually twists the 

 petiole so as to turn the apex of the leaf instead of the sinus toward the 

 center of the plant. N. laydckcri rosea prolifera is a fasciated strain, 

 which stools out around the crown to a remarkable degree, whereas the 

 type makes no lateral buds at all. 



