294 



SUMMARIES OF THE HISTOLOGIC CHARACTERS, ETC. 



polar 10113 both incline to X. bou 



but N. " close us X. giantess, 



I,, i with the raw grains the 



hybrids are as well separated from each other 

 iv from i i i '■• '■ i if gela- 



tinized grains bel ■ like those of N. bo 



while tin- raw grains lean to the other parent ; bnt in 

 the ■ there was no1 found any difference 



in the parental inclinations of both gelatinized and raw 

 grains. The qualitai I ras with the chemical 



trious differences, N. giantess in only two 

 e six reactions inclining to N. bowdeni and in the 

 other four to the other parent; while the other hybrid 

 Ml sis reai tions to N. In the reac- 



tions of chloral hydrate, potassium sulphocyanate, and 

 sodium salicylate N. abundance is closer than X. giantess 

 to A. i in the potassium-sulphocyanate reac- 



the hybrids are clo er to eai b other than to either 

 parent. In the nitric-acid reaction N. giantess is closer 

 to N. i is \ar. corusca major than is N. abundance 



to N. , but the hybrids themselves are very dose. 



In the potassium-iodide reaction .A', giantess leans to 

 N. sarniensis var. corusca major, while the other hybrid 

 incline- to the other parent; but the hybrids are closer 

 other than is either to the parent to which it 

 i lated. The quantitative and quali- 

 tative reactions show most interesting differences and 

 endence. 

 it will be seen by an examination of the preceding 

 table how variable and absolutely unpredictable is the 

 shifting of hybrid properties toward one or the other 

 il inheritance in each of the designa- 

 tions is manifest; but in some instances hybrid and 

 nts are verj closely alike, and in others the hybrids 



more alii ■ more differenl than are the parents, or 



differ more Erom the parents or resemble more 

 ne or the other parent than do the parents them- 

 selves appear to be the same or different. With the first 

 pair of hybrids, .V. dainty maid inclines in the histologic 

 and qualitative reactions, with the exception 

 i i- and arrangement of the lamella 1 , in every 

 on co .V. elegans: while its mate, X. queen of 

 ut two-thirds of the designations 



t. With the second pair. A'. i/iittilcss 



inclini half of the designations to N. bow- 



while N. abundance inclines almost wholly to the 

 parent. With the lasi hybrid, X. glory of swr- 

 the inclination with tl e i ■•■• ption of a single desig- 

 to N. sarniensis var. corusca major. Excess 

 ■licit of development are rarely noted, and no indi- 

 viduality of the hybrid in anj case was recorded. In the 

 ■e reactions there is obvious independence of 

 the qualitative reactions, inasmuch as they may or may 

 not co i. In N. ilit in 1 1/ maid, while in both histo- 



es and qualitai the inclination 



ely to the pollen parent, in the quantitative 



g there is a te i intermediateness, and 



to thi lli parent. I u \ . | here is an 



nation of rds of the histologic proper- 



nd qualits to the pollen parent, while 



in th a leaning 



to the pollen than to the seed parent. 1 d A', giantess 

 one-half of the histologic properties and qualil m re 

 reactions lean to th parent, in the quantitative I 



reactions six of the eight reactions lean to the pollen 

 parent. In X. abundance the histologic properties and 

 qualitative reactions incline almost wholly to the seed 

 parent, in the quantitative reactions six of the eight in- 

 cline to the pollen parent. In N. glunj of sarnia the 

 histologic properties and qualitative reactions incline 

 almost wholly to the seed parent and the quantitative 

 reai tions incline equally to each of the two par. 



Narcissus. (Table C 0.) 



The first two hybrids, while showing throughout the 

 various designations hiparental inheritance, usually bear 

 a (loser relationship to X. poeticus poetarum than to 

 N. poeticus ornatus; and on the whole are closer to one 

 another than to either parent. It is strange that while 

 A', poeticus herrick is in form, hilum, and lamellae closer 

 to N. /mi lint* ornatus than to the other parent, the rela- 

 t ion-hip in size and all other designations is closer to N. 

 poeticus poetarum. N. poeticus dante is in form closer 

 to N. poeticus ornaius, but in all other designations 

 closer to the other parent. In form both hybrids are 

 closer to N. poeticus ornatus, but .A", poeticus I 

 is the closer of the two. In hilum and lamella?, N. poeti- 

 cus herrick shows as close relationship to .A 7 , poi 

 ornatus as does N. poeticus dante to N. poeticus poe- 

 liiniiu. In size, X. poeticus herrick is closer than N. 

 poeticus dante to N. poeticus poetarum. In both polari- 

 scopic figure and selenite reactions both hybrids are 

 closer, and in equal degree, to X. poeticus poetarum. 

 In the iodine reactions the hybrids do not differ and are 

 therefore equally close to A T . poeticus poetarum. 

 Throughout the qualitative chemical reagent designa- 

 tions the hybrids are closer to N. poeticus poetarum. 

 In the chloral-hydrate and nitric-acid reactions X. poeti- 

 cus dante is closer than X. poeticus herrick to .A', poeti- 

 cus poetarum ; but in the chromic-acid and pyrogallic- 

 acid reactions the reverse. Only rare records of deficient 

 development were recorded: in no instance wa- there 

 excess of development or individuality. In the quanti- 

 tative reactions X. poeticus herrick is mid-intermediate 

 or shows a closer relationship to the pollen parent ; while 

 N. poeticus dante is mid-intermediate in three of the 

 seven reactions and shows a closer relationship in two 

 to the seed parent, and in two to the pollen parent. It 

 is of interest to note that while in the qualitative reac- 

 tions both hybrids are throughout very much closer to 

 : In pollen parent than to the seed parent, in the quant na- 

 tive reactions the first leans markedly to the pollen parent 

 and the second to one as much as to the other parent. 



There is seen throughout the designations of the 

 various sets of Narcissi the same swinging of hybrid 

 development to one or the other parent, the independence 

 of each unit-character and unit-character-phase of even- 

 other in its direction and degree of development, the 

 absolute impossibility of forecasting the parental rela- 

 tionship of any designation, and the usually close rela- 

 tionship of the hybrid in its properties, as a whole, to 

 one or the other parent, as is evident in preceding sets. 

 Special features of the Xarcissi group are attached to 

 the relative potencies of certain of the parents that occur 

 in a number of sets, and to the hybrid X. madame de 

 graaff, which in two sets is the pollen parent. X. poeti- 

 cus unittlus is the seed parent in Set 1 and the pollen 

 parent in Sets 2, 3, and 1. As the seed parent, it exhibits 



