338 RONALD J. DINUS 



suggested by Jewell and Mallett (1967) . This conclusion seems especially 

 sound as these were open-pollinated progeny from a rust-free selection in 

 a stand that was 75% infected. Even better field performance can be 

 expected from crossing individuals thus identified. For example, progeny 

 from the cross between 8-7 and 18-27 or its reciprocal averaged only 10% 

 infection in a series of artificial inoculations (Jewell and Mallett, 

 1967) . Consequently, selection and breeding on the basis of artificial 

 inoculation results should yield rapid improvement. 



Judgments based on artificial inoculation results err only on the 

 safe side. Of the progenies resistant to artificial inoculation, none 

 were found susceptible in the field (Fig. 3) even after several years of 

 exposure at two locations. While some potentially useful selections may 

 be overlooked, susceptible ones will not go undetected. 



According to results of the artificial inoculation tests (Fig. 1) , 

 the infected selections, 18-62, 18-40, and 18-41, are undesirable for 

 inclusion in breeding programs. The presence of 18-62 and 18-40 in the 

 upper right quadrant of Fig. 3 supports this conclusion and further 

 demonstrates the close agreement between artificial and field testing. 

 While 18-41 seems undesirable at first glance, its intermediate field 

 performance (Fig. 2) makes outright rejection questionable. Assuming 

 that resistance is controlled by additive genes (Kinloch, 1968) or at 

 least not by a single dominant gene (Jewell and Mallett, 1967), this 

 intermediacy suggests that 18-41 possesses some genes for resistance. 

 That is, it has a larger complement than 18-62 or 18-40, but smaller than 

 8-7. Data from previous artificial inoculations (Jewell and Mallett, 

 1967) support this inference. Progeny from crosses of 8-7 to 18-40 and 

 18-62 were 42 and 44% infected respectively. Crossing 8-7 with 18-41, 

 however, produced progeny which were only 23 percent infected. 



Relying on the artificial testing of open-pollinated progenies would 

 have eliminated 18-41. Thus, such tests may be too severe, but they 

 identify the most resistant selections several years before field tests 

 would. The accumulated data suggest a means of circumventing this weak- 

 ness. For example, partially resistant and potentially useful selections 

 such as 18-41 can be identified by screening control-pollinated progenies 

 or at least progenies of crosses between candidates and specific tester 

 parents like 8-7. 



The rust-free selection, 11-6, performed better in the field than 

 might have been expected on the basis of artificial inoculation alone 

 (Fig. 3). Its performance in other Gulfport plantings has also exceeded 

 expectation. This deviation from agreement infers that 11-6, like 

 18-41, possesses fewer genes for resistance than 8-7 or else has a dif- 

 ferent form of resistance. Available evidence favors the latter premise. 

 First, 11-6 performed quite differently in the two types of test. 

 Second, artificial inoculation of progeny from crosses of 11-6 to 8-7 

 and 18-27 resulted in 53 and 56% infection respectively (Jewell and 

 Mallett, 1967). Progeny from its cross to 18-62 were 95% infected. 

 Thus, 11-6 did not combine as well with resistant selections as 18-41 

 and was highly susceptible in combination with a susceptible selection. 



The field resistance of 11-6 may be dependent upon age or interaction 

 with the environment for its expression. An increase in resistance to 

 blister rust (Cronartium vibicola J.C. Fisch. ex Rabenh.) with age was ob- 

 served by Patton (1961). On the other hand, the progeny of 11-6 tended to 



