81^ 



July and the second in the last two weeks cf August, this 

 sta'^ement being true for all the stations considered except- 

 ing Oakland, fcr which station they both occur relatively 

 early in the season, in the last two^aeks in Jiine and July 

 respectively. A tiiird feature which is common to mo stf though 

 not all) of these graphs lies in the fact that the up^irard 

 slope is more gradual (before the occurrence of the high 

 midsummer maxima) than is the downward slope fafter the occur- 

 rence of the maxiiTH.) A generalized tem-oerature efficiency 

 graph representing averages of the corredponding values for all 

 of the sta'''io'.s is n;t symmetrical about the ordinate for its 

 highest midsummer value; it slopes uow^-rd less rapidly than 

 down7/ard. A foutth general characteristic of these graphs 

 lies in the fact that the final low index values of the frost- 

 less season are not ver«) different for the various sta'^ions. 

 The following considerr.tion of the graphs for some of the in- 

 dividual stations vlll serve to bring out the points mentioned 

 above and will give opportunity to note exceptions to the 

 generalized statements just made. 



In regard to the forms and other characteristics of the 

 temperature efficiency graphs, the nine stations studied may 

 be grouped into five classes: fl) Ohewsville and Monrovia, 

 fEi Baltimore, Darlington and Coleman, (3) Easton and 

 Princess Anne, f4) College, and (5) Oakland. The last two 

 stations i'o not appear to fit jnto any of the first three 

 classes and they ar: not aliky au ':':i'^^ ma-jt aot be regarded as 

 representing separa/'e classes. These five groups are dis- 

 cussed in order below. It will be noted that the stations 

 of groups 1, 2, and 3 are located near each other, and this 



