16 



INTRODUCTION. 



both parent forms. My Nicotiana X N. paniculata had 

 in the second and third generations mostly much nar- 

 rower leaves than in the parent species. 



6. The sterility and inconstancy of the offspring of 

 hybrids has often misled botanists into conclusions which 

 are not supported by experience. As may be seen by the 

 facts already set forth, it is absolutely incorrect if it is 

 concluded that all hybrids must necessarily die out 

 quickly because of the many and various properties which 

 are combined in them. The variable forms resulting 

 from a crossing are the material from which not only 

 gardeners produce their new varieties, but which are also 

 biologically valuable in that they furnish new species 

 in the economy of nature. 



(c) Back-crossings of Hybrids with Parent Species 

 (A 9 X B4 ) 9XA,J,(A9XB,J) 9 X B^, A9 

 X (A X B) $ . As long as great stress was laid on the 

 role which the pollen of the seed-parent species played 

 in the production of a hybrid a careful distinction was 

 made that advancing hybrid forms approached the male 

 parent species and reverting hybrid forms approached the 

 female species. These distinctions are, however, accord- 

 ing to the mass of recent experiments, of very secondary 

 or of no significance. 



On fertilization of a hybrid with parental pollen 

 there appear, as a rule, a moderately variable progeny. 

 Intermediate forms between the hybrid and the parent 

 are the most numerous and most fertile. With these are 

 a smaller number of individuals which are similar to the 

 primary hybrid or to the parent species, and both kinds 

 are usually of lessened fertility. 



The three-fourths hybrid (A X B) 9 X A $ are 

 often moderately fertile with their own pollen and seem 

 to produce stable races more readily than the primary 

 hybrid, as in /Egilops speltcvformis. Gartner noted 

 many times that in later generations of three-fourths 

 hybrids the pollen was nearer normal and the fertility 

 greater, as in Dianthus (chinensis X barbatus) X D. 

 barbatus, and also in other three-fourths hybrids of Dian- 

 thus, Lavatera, and Nicotiana. 



If the three-fourths hybrid (A X B) 9 X A 3 be 

 fertilized with the pollen of A, there will be produced a 

 seven-eighths hybrid or the third hybrid generation 

 which, as a rule, is very similar to the parent species 

 represented as seven-eighths of the product, but which, in 

 individual specimens, still shows material differences in 

 form and fertility. The last trace of the one original 

 parent species is obliterated in the fourth, fifth, or even in 

 the sixth hybrid generation. 



Kb'lreuter and Gartner have effected the transforma- 

 tion from one parent species to the other in many in- 

 stances. They found that for the transformation to be 

 complete three to six generations are required, usually 

 four to five. Manifestly, the greater or lesser duration 

 of the period of transformation depends in part on col- 

 lateral conditions. Godron found that Melandryum 

 album X M. rubrum fertilized with its own pollen re- 

 verts in the second generation to the parent species, 

 while Gartner considered three to four generations neces- 

 sary to carry one species over to the other through fer- 

 tilization with parental pollen. 



In general, the products of the fertilization of one 

 parent species with hybrid pollen, as A 9 X (A X B) $ , 



are similar to those of the reverse fertilization, but 

 observers agree that the variety of forms is greater if 

 the hybrid is used as the male factor, as in Dianthus 

 and Salix. 



As in the direct progeny, so also in back-crossings 

 of hybrids, new properties frequently appear which are 

 absent in the present forms, but which are often found 

 in related species or races. 



Hybrids of Several Species. Triple Hybrids. Kol- 

 reuter, during the first year of his research, succeeded in 

 combining three entirely different Nicotiana species in 

 one hybrid form. The only formulas according to which 

 such a combination can be made are: (A X B) 9 X 

 C <?, C ? X (A X B) <J and (A X B) 9 X (A X C) * . 

 In the genera Dianthus, Pelargonium, Begonia, 

 Rhododendron, Nicotiana, Achimenes, Calceolaria, Salix, 

 Hippeastrum, Gladiolus, and several others, many 

 such combinations have been produced without 

 especial difficulty. Differentiation must be made be- 

 tween combinations of three entirely different species, 

 and combinations in which two or all three of the factors 

 are closely related. There are several manifestly different 

 species which in hybridization with one another art 

 almost like races of the same species, as Melandryum 

 album and M. rubrum; Vitis vinifera, V. cordifolia, 

 V. cestivalis and V. labrusca; Lobelia fnlgens, L. splen- 

 dens and L. cardinalis; Rhododendron ponticnm, R. 

 arboreum and R. catawbiense ; Rhododendron flamim, 

 R. viscosum, R. nudiflorum and R. calendulaceum ; Ber- 

 beris aquifolium and nearly related species. 



Hybrids produced by crossing the hybrids of two spe- 

 cies of these groups with a third species of the same 

 genus can as little be considered true triple hybrids as 

 hybrids of three of the narrow groups belonging to the 

 Vitis, Lobelia, and Rhododendron species. True triple 

 hybrids formed from three essentially separate species 

 usually produce a moderate variety of forms, especially 

 if the male parent is a hybrid. On the other hand, in the 

 combination which is easiest to produce, and which is 

 formed on the formula ( A X B) 9 X C $ , the type of C 

 usually predominates, as in Nicotiana (N. rustica X N. 

 paniculata) 9 X N. langsdorffi $ , Achim.enes. (A. 

 grandiflora X A.. Candida) 9 X A. longiflora $ , and 

 several of the Gesneracece. 



The hybrids of Erica when crossed produce as uni- 

 form a progeny as do the pure species. Several Salix 

 hybrids behave in a similar manner. 



Triple hybrids in many genera (Pelargonium, Be- 

 gonia, Rhododendron, Achimenes, Isoloma, Cypripe- 

 dium, Gladiolus) are for these reasons very valuable to 

 gardeners. If they produce seed their progeny are very 

 unstable. 



Hybrids of Four to Si.r Species. If the hybrids be- 

 tween very nearly related species (Vitis, Rhododendron, 

 and so forth) are not considered, hybrids from four or 

 more parent forms are moderately rare. They are found 

 especially in the genera Dianthus, Pelargonium, Bego- 

 nia, Rhododendron, Nicotiana, Salix, Hippeastrum, and 

 Gladiolus. The artificial combination of different species 

 in a single hybrid form was practised to the widest ex- 

 tent by Wichura, who has combined in Sali.r six species. 



Hybrids of Combined Hybrid Offspring. In sev- 

 eral genera (Pelargonium, Fiifhsia, Begonia, Rosa, 



