plant is too low for practical purposes," and explains how reproductive vigor is 

 restored in plants from single and double crossing of two or four inbred lines. 

 Actually the maize breeders also selected among their inbred lines for reproductive 

 vigor, and the same principle can be applied to western white pine and other conifers. 



An example of the effect of thus selecting for comparatively high self-fertility 

 can be drawn from our results with western white pine. If we had continued the inbreed- 

 ing program using only those highly self-fertile parents that yielded 60 or more filled 

 Sj seed per cone (Bingham and Squillace 1955, table 1), we would have selected seven 

 of the nine Si lines in which second- generation inbreeding gave low to fair, but 

 consistent yields of filled S2 seed (3 to 34 per cone) . 



The effectiveness of such selection for reproductive vigor is emphasized when, as 



in table 6, a comparison of second-generation cone and seed yields is made between 



trees from a comparatively self-sterile parent (tree 19) and those from comparatively 

 self-fertile parents (trees 25 and 58) . 



Snyder (1968) shows that in Mississippi slash pine yield filled seed per cone 

 and germination of S^ seeds are lower than in most other pines. He also shows (1968 

 and 1972) that inbreeding depression--measured in height growth reduction--increases in 

 slash pine S^ trees between the first and fifth years of growtli (cf. also Bingham and 

 Squillace 1955, and Barnes 1964, for the same phenomenon, ages 1 to 12 years, observed 

 in western white pine). Nevertheless, Snyder's cones from "self-compatible" parents 

 yielded 2 to 13 filled seed per cone compared to 30+ seed per wind-pollinated cone, 

 and one-third or more of the S^ seeds germinated. 



Snyder (1968) concluded, from the low filled-seed and seedling yields per 100 self- 

 pollinated strobili, from the probable (but unknown) delav in flowering of the Sj trees, 

 and from apparent indications of a relatively low level of nonadditive variance in most 

 traits of southern pines, that an inbreeding program was neither feasible nor practical 

 for improving most characters of slash pine. His own evidence for the first inbred slash 

 pine generation, however, coupled with evidence from second-generation crossing in four 

 other species in Pinaaeae , makes these conclusions seem premature. The arguments of 

 Franklin (1969) against inbreeding for improvement of loblolly pine {Pinus taeda L.) are 

 much stronger, including unknown effects of probably greater pregerminative selection 

 in inbreds, hypersensitivity of inbreds to environmental stress, the relative difficulty 

 (thus high cost) of maintaining inbreds, and the present disaffection of even the maize 

 breeders with the traditional inbreeding-outcrossing hybrid method. 



Despite these problems, continuation of experimental inbreeding programs in some 

 conifers seems warranted. Consistent, if small, Sj and S2 seed yields, and S^ plants 

 that fruit within reasonable periods, have been obtained in at least four conifers. 

 Also, experimental efficiency probably could be greatly enhanced through selection of 

 comparatively self-fertile parents for subsequent cycles of inbreeding. 



Questions that remain for the inbreeder center on long-term economic and genetic 

 practicability, and on the stability of gains under inbreeding as opposed to outcross 

 breeding. To answer these questions, we need to proceed on an experimental basis 

 toward answering such subordinate questions as: What is the gain from one generation of 

 inbreeding followed immediately by single crossing of selected, vigorous Si's? and, 

 Is reproductive vigor of the inbreds restored in these early single-cross lines, so that 

 the double crosses may be produced quickly, economically, and in quantity? 



It would seem that gains anywhere near those achieved in maize double crossing 

 would finance such an abbreviated inbreeding program in trees. If growth of single- 

 cross lines from Si crossing represents little or no gain from heterosis, and bears no 

 relation to vigor of parental Si lines, then it will be time to reconsider the practi- 

 cality of inbreeding programs in forest trees. 



14 



