in each of the 5 years was adjusted by the ratio of individual area mean to the all- 

 areas mean; if the interaction between weather and site effects is minimal, this adjust- 

 ment should effectively discount inherent differences in site productivity. Adjusted 

 productions of total graminoids, of total forbs, and of total vegetation were then 

 evaluated against selected weather factors. 



Various expressions of precipitation, air and soil temperatures, solar radiation, 

 wind, and soil moisture status in different months were examined for possible linear 

 relationships to herbage production. Fifty-three such independent variables were 

 examined. From these, I selected the 16 shown in table 6 as possibly being related 

 singly or in combination to the amount of herbage produced every year. The measurement 

 units for these 16 independent variables are: 



Soil moisture -- logarithm of ohms electrical resistance; 



Precipitation -- accumulated inches for given period: 



Solar radiation average cal . /cm. ^/day for given period: 



Air temperature -- daily maximum °F. averaged for given period; 



-- daily minimum °F. averaged for given period; 

 -- degree-hours (the daily average of hours the 



temperature remained within 5° F. of the extreme, 



multiplied by the extreme); 

 Soil temperature -- same measurement units as air temperature, but 



measured on a shaded soil surface; 

 Wind velocity -- average miles per hour for given period. 



Few of these variables were independently correlated significantly with the produc- 

 tion of either graminoids, forbs, or total vegetation in this limited test. Conse- 

 quently, the significant positive correlation of production with May precipitation and 

 the negative correlation with the maximum temperatures of the shaded soil surface are 

 particularly conspicuous (table 6). Earlier, I observed that total production on 

 similar mountain grasslands may be related more closely to precipitation falling during 

 the period of most active growth than to precipitation at other times (Mueggler 1967). 

 Smoliak (1956) found production on the short-grass prairies of Alberta to be highly 

 correlated (r = 0.86) with May-plus-June precipitation, and significantly negatively 

 correlated (r = -0.53) with seasonal mean temperatures. Failure to verify suspected 

 correlations with soil moisture at the beginning of plant growth might be explained by 

 the small variation in soil moisture between years. Soils on these grasslands usually 

 are near their maximum moisture-holding capacity, down to a depth of 50 cm. at least, 

 at the beginning of plant growth (Mueggler 1971). Ordinarily, this is also true down 

 to 100-cm. depths on northeast exposures, but not on southwest exposures. The negative 

 correlation with maximum temperatures of the shaded soil surface and positive correla- 

 tion with minimum air temperature (degree-hours) suggests that growth of these mountain 

 grassland plants may be favored somewhat by rather cool weather in June and July. 



I tested various combinations of independent variables by multiple regression 

 analyses to determine the combination that would best predict herbage yield. A combina- 

 tion of the five following weather measurements gave the highly significant coefficient 

 of multiple determination (R^) of 0.86 when correlated with total herbage yield: X3, 

 May precipitation; X7, maximum air temperature (°F.) in July; Xq, maximum air temperature 

 (degree-hours) in June; X^^, minimum air temperature (degree-hours) in June; and X^^, 

 maximum temperature (degree-hours) of shaded soil surface in June. The equation used 

 to estimate total herbage production (Y^) has a standard error of estimate of 

 83 lb. /acre (mean production = 1,030 lb. /acre): 



Yi = 1,194 + 169X3 - 19.2X7 - 2.31X9 + 2.40X11 + 3.50Xm. 



16 



