596 HENRY D. GERHOLD 



Clearly, the best answer will be found somewhere between these two 

 extremes. Predicted and actual genetic gains will depend on valid genetic 

 information. Yet the data must be derived from the parent genotypes which 

 should be selected on the basis of this same information. This dilemma 

 exists in every dynamic breeding scheme because the parents and genetic 

 parameters change in subsequent generations. The breeder aims for the 

 ideal solution through a series of approximations, using a judicious 

 mixture of hard facts, experienced judgment, and faith. Resources must 

 be committed at the outset both to finding superior genotypes and to 

 obtaining genetic information. Neither element can be sacrificed com- 

 pletely, but the relative amount of effort devoted to each may be shifted 

 toward the greater need. 



In my opinion the greater need in the situation that we have been 

 examining is for better definition of genetic parameters involving the 

 two pathogens. My recommendation is to resolve the difficulties in two 

 steps by dividing the ultimate improvement goal into two intermediate 

 goals. In other words, genes that confer resistance to blister rust 

 would be selected separately from those that confer resistance to white 

 pine weevil, thereby reserving the more complex interactions for later 

 generations. Accordingly, the following suggestions are offered for 

 the first generation: 



1. Choose two sets of parents, one for blister rust resistance and 

 the other for weevil resistance. Maximize selection intensity for resis- 

 tance in each, merely avoiding negative selection for other yield-related 

 traits. 



2. Restrict matings to within each set of parents. 



3. Test resistance in two sets of environments, protecting each 

 set against the other pathogen. Test each set of parents for resistance 

 only against the pathogen considered in its selection. 



4. Subdivide each family for testing, exposing a larger portion to 

 the pathogen in the nursery, and comparing survivors with the unexposed 

 portion in forest plantings to determine the validity of nursery selec- 

 tions. Take measurements at about 4-year intervals to define growth and 

 mortality trends and their interrelations. 



5. Obtain the data needed for two independent selection indices, 

 each containing parameters of growth and of resistance to one or the 

 other pathogen. Compared to simpler selection methods, index selection 

 offers greater efficiency and discipline in accounting for genetic and 

 economic factors that influence genetic gains and their values. 



During the first generation a firm foundation would be laid for 

 further progress, possibly sacrificing some genetic gain in favor of 

 lower risk and greater flexibility. A decision on combining both types 

 of resistance into a single variety would be postponed until some of the 

 uncertainties have been clarified. In the meantime two improved varieties, 

 each with one type of resistance, could be produced and could be used 

 silviculturally, singly or in mixture, in conjunction with other measures 

 for controlling blister rust or weevils. 



