68 



mar^h better than the four-v.-efek graphs. Thip fact will be 

 brought out by a brief reference to them at this point. 



The values of the temperature indices for the four-week 

 periods show the seasonal mar-^hes of this condition for the 

 various stations, from low values in May to high midsumner 

 values and then to lov: vmlues again in the last part of the 

 season. The grapn.s for all of the stations except Oakland 

 show a steeper slope after the midsummer maximum has been 

 passed than for the periods during which the temperature was 

 rising to this maximum. The two maxima which were present 

 in most of the two-week graphs ar^ eliminated in the four- 

 week averages and the graphs of temperature values show in- 

 stead a period of about six weeks during which this condi- 

 tion remains approximately constant. 



The four-week evaporation and light data, show the gener- 

 al characteristics of the seasonal marches of these -conditions 

 previously noted as exhibited by the two-week data. It will 

 be seen in the first place, that both graphs exhibit a downward 

 slope from the beginning to the enJ of the season; and, in 

 the second place, that both graphs show, in addition to their 

 high primary maximum in the early part of the season, one or 

 more secondary maxima later. In some ciases the secondary 

 maxima in the evaporation graphs coincide with temperature 

 maxima. Both of these general characteristics shorrn in the 

 four-week graphs of evaporation and light are shown by the 

 two-week graphs but since small variations are eliminated by 

 averaging the overlapping periods, there are fewer secondary 

 maxima in the four-week graphs. In the case of evaporation, 

 there is usually one secondary maximum occurring in or near 



