20 THE WATER-BALANCE OF SUCCULENT PLANTS. 



Observations of the internal temperature of a sahuaro trunk were made 

 during: the period April 12 to May 6, 1907, by means of bent thermometers, 

 the bulbs of which were inserted about 3 inches below the surface, one on 

 the north and the other on the south side. As shown by table 3, the aver- 

 age temperature indicated by the latter was l.l" higher in the morning- 

 and 4.6° in the afternoon than that shown by the former. 



Observations on diurnal change of internal temperature were not made 

 during the time the sahuaro was in flower; but in June and again in No- 

 vember observations were carried through 24 hours. In each case a bent 

 thermometer was inserted directly below the point where the first buds ap- 

 peared (southeast) and another directly opposite to it (northwest). Table 

 4 and curves (fig. 11) show the results. 



The average temperature was in both cases higher on the southeast and 

 the period of greatest difference in June lay between 5a.m. and 2p.m. It 

 seems fair to infer that had the observations been made in April and May, 

 when the sun was farther to the south, the difference between the two sides 

 would have been still more marked. Naturally in November the plants 

 were less warmed by the sun's rays, and consequently internal tempera- 

 ture showed less variation. 



While these observations are too limited to admit of further discussion, it 

 is evident that the tissues on the south and southeast sides show a greater 

 number of heat -units than those on the north and northwest, and since the 

 sum of heat-units is a factor of the first importance in determining the date 

 of flowering, it may well be in this case that the regular order in which 

 the buds appear and the flowers open is determined by the same factor. 



INFLUENCE OF CHANGES OF AIR-TEMPERATURE. 



Of necessity insolation can hardly be considered apart from temperature 

 changes induced by it in the tissues of the sahuaro, but in the discussion 

 thus far it has been connected chiefly with transpiration, since it is proba- 

 bly the increased transpiration of that part of the trunk subjected to the 

 strongest insolation that causes the greater contraction of that part. The 

 effect of seasonal changes of air-temperature would be quite different, and 

 whatever contraction or expansion might arise from this cause would nat- 

 urally be very difficult to differentiate from that caused by variations in 

 amount of available soil-water; but under certain conditions this can be done 

 to some extent. 



An example of expansion following rise of temperature is presented by 

 the records from November 2 to 16, 1906. During this period there was 

 continued high temperature, but no rain had fallen since August 20, and 

 the intervals had been contracting pretty steadily; but the measurements 

 of November 3,7, and 14 showed distinct expansion of 22 intervals. This 

 is perhaps the most conclusive case of the kind, but the following furnishes 

 additional evidence of the same sort: Between March 16 and 22, 1907, a 



