Examining causes for mortality provides insight into the problem of modeling the combined 

 uneven- and even-aged mortality data. For the uneven-aged data, the principal causes of 

 mortality were lightning, wind, bark beetles, and mistletoe (Pearson 1939; Wadsworth and 

 Pearson 1943; Myers and Martin 1963). The amount would indicate that these are noncatastrophic 

 forms of mortality. The main causes for mortality in the even-aged plots were snowbreak and 

 rust (Schubert 1971, 1974). Of the 160 trees that died on those subplots used in equation de- 

 velopment, 86 percent died because of snowbreak in the first growth period after cutting, and 

 12 percent died of rust in the second period. 



A problem exists as to whether snowbreak should be considered noncatastrophic or catastrophic 

 mortality. Accelerated mortality after cutting is not unique to the Taylor Woods study. Also, 

 windthrow after cutting has been widely reported (Alexander 1973 and 1975; Reukema 1970; 

 Reukema and Pienaar 1973; and Williamson and Price 1971) while both Boldt (1970) and Williamson 

 (1976) reported serious problems with snowbreak. Barrett (1965) also noted that "extensive 

 precautions were taken to protect the stand from damage by porcupines and snowbend, the two 

 greatest threats to maintaining the designated tree densities." The problem of snowbreak and 

 bending has been described as being common to the Southwest, particularly among young dense 

 stands such as the Taylor Woods subplots (Pearson 1950 and Schubert 1974) . 



Should, therefore, a mortality agent described as common be considered as a natural 

 result of thinning? The answer is subjective and cannot be presented satisfactorily in this 

 study. To examine how the inclusion of the Taylor Woods mortality data affects the prediction 

 of uneven-aged mortality, the chi-square goodness-of-f it tests were made to the even- and 

 uneven-aged mortality equations using just the uneven-aged data. The resulting chi-square 

 statistics indicated that the inclusion of the even-aged data has deteriorated the predictive 

 capability for the uneven-aged plots. Whether this result indicates that snowbreak is non- 

 normal (that is, catastrophic) mortality is not clear. A comparison of the goodness-of-fit 

 tests across diameter class (tables 29 and 30) to those for the uneven-aged equations (tables 

 11 and 12) indicates that the greatest differences occur in the 4- through 6-inch diameter 

 classes, and these are the classes where most of the Taylor Woods data occurred. Despite 

 these problems, I decided to also test these two equations in the validation phase along with 

 the uneven-aged blackjack and yellow pine equations. Perhaps the equations developed with the 

 even-aged data will prove to be better predictors over the long run. 



72 



