142 Clyde E. Leighty 



Comparing first the means for average yield of culm per plant in deci- 

 grams, it is seen that these are, respectively, 3.433 ± .096 for series 1219; 

 7.890 ± .118 for series 1221; 3.950 ± .102 for series 1257; and 9.087 ± 

 .124 for series 1259. It is thus clear that both varieties agree in having 

 a greater mean for yield of culm when the plants are grown in hills than 

 when grown in drills. The standard deviations are likewise greater for 

 the plants gro-^ni in hills. There is, then, a larger yield per culm and a 

 greater variability in yield per culm when plants are gro\\Ti in hills than 

 when they are gro"wn m drills. 



In the average height of plant in centimeters, the means are 55.850 ± 

 .432 for series 1219; 58.683 ± .290 for series 1221; 49.566 ± .416 for series 

 1257; and 51.916 i .252 for series 1259. The plants grown in hills, there- 

 fore, are on the average slightly taller than those growm in drills. As 

 regards the standard deviations, conditions are reversed from what they 

 were in average yield. For series 1219 the standard deviation is 11.114 

 ± .306, for series 1221 it is 7.462 ± .205, for series 1257 it is 10.703 ± 

 .294, and for series 1259 it is 6.486 ± .178. It is thus seen that there is 

 greater variability for average height m plants growm in drills than in 

 plants gro'wn m hills. 



The means for average number of kernels per culm of plant are about 

 twice as high in series 1221 and series 1259 as in series 1219 and series 

 1257, respectively. The standard deviations are slightly greater, also, for 

 series 1221 and series 1259. These results indicate that culms grown on 

 plants in hills produce nearh^ twice as many kernels as do culms growm 

 on plants in drills, but that the variation is about the same. 



The means for average number of spikelets per culm of plant are 

 about twice as high in series 1221 and series 1259 as in series 1219 and 

 series 1257, respectively. The standard deviations also are slightly 

 greater in the two former series. This indicates that there are about 

 twice as many spikelets produced on culms of plants gro^^^l in hills as on 

 those gro^^^l in drills, but that the variation m number of spikelets on such 

 culms is only slightly greater. 



For average weight of straw per culm of plant in decigrams, the means 

 for series 1221 and series 1259 are again nearly twice as high as for series 

 1219 and series 1257, respectively, and again the standard deviations are 

 greater for the two former series. About twice as much straw per cuhn 



