370 



SCIENCE 



[N. S. Vol. LIII. No. 1372 



components of the doublet (in tlie case of the 

 hundred oi' so members which can be re- 

 solved) are of exnctl.y eqiial intensity, and 

 therefore it seems safe to assume that the \n\- 

 rcsolved doublets are at least symmetrical, 

 and to use them. For 3873.504, Grebe and 

 Baehcm obtained a shift of 0.58 km. (average 

 of five consistent determinations from differ- 

 ent plates), and Schwarzsehild 0.45 km. 

 (average of fo\ir consistent plates). No other 

 investigators have used this line. For 3858.- 

 822, Grebe and Bachem obtained 0.42 km. 

 (average of six consistent determinations), 

 St. John 0.40 (average of four different 

 methods, of which only tlie first two are 

 wholly independent and so entitled to the 

 most weight, these two, a and h, yielding 

 0.46 km.), Schwarzsc'hild 0.39 km. (average of 

 four readings — three consistent). 



Using the 0.46 km. value of St. John, these 

 five determinations for the two lines average 

 0.46 km. /see. In all cases this is the shift 

 between lines radiated by the center of the 

 sun, and by the arc. But St. John Hoc. cit.) 

 and Adams have both obtained reliable evi- 

 dejiee that at the center of the sun there is 

 a rising current of about 0.12 km./sec, com- 

 pared to the rim.i" This tends to mask 

 the Einstein effect. The true value of this 

 effect, as experimentally determined, is then 

 0.46 + 0.12 — 0.58 km./sec, compared to the 

 theoretical 0.634. Wliile the data are far too 

 meager to draw any final conclusions, it is 

 wortliy of notice that tJie results of all ob- 

 servers are truly consistent on reaUy good 

 linos. The groat diseropanc.y between St. 

 John's and Grcbo and Bachem's general 

 averages has been the puzzling factor, tlms 

 far. The author believes that he has a partial 

 explanation for this, and will present it in a 

 later paper, together with a list of lines which 

 are suitable for use, as far as condition (2) is 

 concerned. 



It might be added that, for the nine lines 

 quoted by Grebe and Bachem''" » (X 3858.822 



10 SeliwuTz-sciiild 's results indiomte a Mlmg cur- 

 rent, not a rising, as quoted by Grebe and Bnohem, 

 but are too discordaait to have any vahie. St. 

 John's are very reliable. 



and A 3851.427 being accidentally omitted), 

 the agreement among different observers is 

 worse than indicated, due to Grebe and 

 Bachem's consistent misquoting of St. John's 

 results, as well as other errors. The correct 

 averages are: G. and B. 0.57, Schwarzschild 

 0.63, St. John 0.17 (or 0.26 using methods 

 a and h only), Everslied and Royds 0.67. 

 General weighted average 0.50, or 0.52, using 

 0.26 for St. John. 



If all eleven lines are used, the averages 

 become: G. and B. 0.52 (eleven lines), 

 Schwarzschild 0.57 (nine lines), St. John 

 0.22 (eight lines, or 0.30, two good metliods 

 only), Evei-shed and Eoyds 0.67 (two lines). 

 Average (weighted according to the number 

 of lines), 0.46, or 0.48, using 0.30 for St. 

 John. To all these values should be added 

 0.12 km. to obtain the true rim — arc value. 



It should also he added that, in the author's 

 opinion, St. John's method (c), and Grebe 

 and Bachem's recent calculation^ of 100 ON 

 lines add comparatively little weight to the 

 argument, as they involve tlie use of How- 

 land's standards. Since Rowland used both 

 terrestrial and solar wavo-lengtJis, in obtain- 

 ing his table of standard lines, the Einstein 

 shift (if real) is hopelessly involved in the 

 measurements and can not be definitely ex- 

 tricated by any such method as that recently 

 used b,v Grebe and Bachem. 



Raymond T. Birge 

 University op Caliporni.\ 



a new high temperature record for 



GROWTH 



A RECORD of growth of young joints of a 

 prickly pear (Opuntia) at 50° C. and 51.5° 0., 

 and of the active elongation of etiolated stems 

 of the same plant growing at 49° C. was pub- 

 lished in 1917. Previously to that time Dr. 

 J. M. McGee had foimd that the mature 

 .ioints of the same Opuntia might reach tem- 

 peratures of 55° C. in the open without 

 damage, which was a record for endurance of 

 the higher plants in air. 



In the repetition of the growth measure- 

 ments at the Desert Laboratoi-y late in March, 

 1921, young .ioints which might reach tern- 



