6 



INTRODUCTION. 



detect microscopically any blending of patent characters, 

 even though these may occur. Some may demur to 

 accepting conclusions drawn from comparison of the 

 hybrids of two parents that are even moderately removed 

 from each other in affinity, particularly since we know 

 that such are frequently less fertile than the pure product 

 of either parents, or are entirely sterile. The objection 

 will afterwards be considered, but here I may premise 

 that, as a rule, whether the parents are remotely or closely 

 related their evenly blended peculiarities appear, if com- 

 parison is at all possible. 



" To the above general conclusion, however, we must 

 make an important exception. In not a few cases, which 

 will afterwards be cited, a separation or prepotency of 

 the sexual molecules of each parent seems clearly to be 

 indicated. 



5. RELATIVE STABILITY or PABENT FORMS. 



" Some species, both in the wild state and under culti- 

 vation, show a greater degree of stability, or want of 

 variation tendencies, than do others. This is probably to 

 be explained by an average structure having been slowly 

 but steadily evolved through crossing and recrossing of 

 an aggregate of like individuals with survival of those 

 best fitted for a set of environmental conditions that re- 

 mained constant through long periods of time. These, 

 therefore, even when removed to rather disadvantageous 

 surroundings, do not readily exhibit change. As exam- 

 ples, I may name Erica tetralix, E. cinerea, and Philesia 

 buxifolia. One finds that the opposite is equally true 

 of not a few species. Thus, if a series of individuals 

 of Qeum rivale or Dianthus barbatus (cultivated) be 

 compared microscopically, considerable variation is 

 traceable. 



" But even species which are considered to vary little, 

 if compared from wide areas, may present unexpected 

 changes. An interesting illustration is furnished by a 

 plant just cited as one of the most invariable, viz. Erica 

 tetralix. I have shown elsewhere * that this species re- 

 solves itself into four subspecies, three of which are 

 found in Connemara, and these, so far as they have been 

 experimented on, remain true under cultivation. It is 

 necessary, therefore, in the selection of a hybrid to 

 know the exact type of each parent, if not the actual 

 parent, and to examine such alongside the hybrid 

 offspring." 



Macfarlane made detailed studies of the microscopic 

 peculiarities of nine sets of parent-stocks and hybrid- 

 stocks, including the following: 



1. Lalageria rosea, Phileaia buxifolia, P. veitchii 



2. DianthuB alpinus, D. barbatus, D. grievei. 



3. Geum rivale, O. urban mn, G. intermedium. 



4. Ribe growularia, R. nigrum, It culverwcllli. 

 6. Saxifraga geum, 8. aizoon, 8. andrewgii. 



6. Erica tetralix, E. ciliarU, E. wateoni. 



7. Mensiesia empetriformis, Rhododendron chamecistiu Brv- 

 anthus erectus. * 



8. Masdevallia amaJiilin, M. veitchiana, M. ehelsoni 

 . Cypripedium pioerianum, C. insigne, C. leeanum. 



He also recorded many data respecting other hybrids 

 and parents, including in the text only some special 

 features which seemed to deserve consideration, to- 

 *Trni. Bot. 800. Edln., xnc. 1891. 



gether with a rather full account of the characters of a 

 graft hybrid, Cytisus adami. The following is Mac- 

 farlane's " General Summary of Kesults on Seed 

 Hybrids " : 



" It has been demonstrated that in hair production, 

 if the parents possess one or more kinds that are funda- 

 mentally similar, but which differ in size, number, and 

 position, the hybrid reproduces these in an intermediate 

 way. Illustrations of this were presented by Qeum inter- 

 medium, Erica watsoni, Cypripedium leeanum, and Mas- 

 devallia ehelsoni. But if only one parent possesses hairs 

 over a given region the hybrid usually inherits these to 

 half the extent, as in the petals of Dianthus barbatus 

 and some floral parts of Bryanthus erectus. If the hairs 

 of two parents are pretty dissimilar, instead of blending 

 of these in one, the hybrid reproduces each, though re- 

 duced in size and number by half. The gland hairs of 

 Saxifraga andrewsii, the simple and gland hairs of Ribes 

 culverwellii, and those on the vegetative organs of Bryan- 

 thus erectus are examples. The peculiar case of air dis- 

 tribution in relation to color formation noticed in the 

 sepal of Cypripedium leeanum may also be noted here. 



" In the formation of nectaries as traced in Phila- 

 geria, Dianthus, Saxifraga, Ribes, etc., the above prin- 

 ciples also hold. 



" The distribution of stomata over any epidermal 

 area has been proved to be a mean between the extremes 

 of the parents, if the stomata of the parents occur over 

 one surface or both, and if the leaves are similar in 

 consistence, but, as in Hedychium sadlerianum, and to a 

 less degree in Saxifraga andrewsii, if the stomatic distri- 

 bution and leaf consistence differ in the parents, this may 

 give rise to correspondingly different results in the 

 hybrid. 



" In amount of cuticular deposit, and arrangement 

 of it into ridges or other localized growths, hybrids have 

 been proved intermediate between the parents. We may 

 merely recall here the case of Philageria stem, which in- 

 herited cuticular ridges from Lapageria, though reduced 

 to half the size, since the Philesia parent was devoid of 

 them. 



" As Wichura has already proved for the vegetative 

 leaves of hybrid willows, the venation of hybrid leaves is 

 very uniformly intermediate between those of the parents. 

 Figures are given with this paper of the vegetative leaves 

 of Philageria and Saxifraga, and of the petals of Dian- 

 thus and Geum. The relation of the bundles to special 

 terminations, as in the water stomata of Saxifraga, is in 

 conformity with the venation. 



" But the growth of tissue in a hybrid which is to 

 determine the outline or angular position which any 

 organ or part of one will assume is intermediate between 

 those of the parents when the latter show traceable dif- 

 ferences. Thus the sepals and petals, as also the styles 

 and style-arms, of Qeum intermedium, the floral parts 

 as a whole of Saxifraga andrewsii and Ribes culverwellii, 

 the frilling of some of the floral parts of Bryanthus and 

 Cypripedium leeanum are pronounced cases, while minor 

 ones have been referred to. 



" Turning to minuter anatomical details, every hy- 

 brid has yielded a large series of examples which prove 

 that the size, outline, amount of thickening, and local- 

 ization of growth of cell walls, is, as a rule, intermediate 



