750 



SCIENCE 



[N. S. Vol. XXXIII. No. 854 



on Kennard while the wind was blowing from the 

 slope across the line were very unsatisfactory. 

 When observations were made over the line with 

 the wind blowing across it towards the slope of 

 the hill, they were of the required degree of accu- 

 racy. The total range in the values for the several 

 observing periods for this direction was about 8" 

 of are. During each observing period the range 

 of the values for the sixteen pointings was small. 

 It is believed that the air blown from the hill 

 across the line was of a different temperature and 

 was the cause of the lateral refraction. 



The large errors in the observed directions of 

 primary triangulation seem to be due to three 

 principal causes. First, to the asymmetrical image 

 of the light or heliotrope caused by wind when the 

 line is low. Second, to lateral refraction, caused 

 by a line passing close to a hillside or mountain- 

 side with the wind blowing from the slope across 

 the line. Third, to the very unsteady lights when 

 the instrument is low with the line passing close 

 to the ground near the station. The first two 

 causes produce constant errors, that is, each of 

 the 32 measures is affected in the same direction. 

 The last cause makes large accidental errors. 



The Modern Potentiometer : Dr. W. P. White, of 



the Geophysical Laboratory of the Carnegie 



Institution of Washington. 



For many purposes it is desirable to avoid the 

 slide wire. The construction of potentiometers of 

 wide range in which only switches are used for- 

 merly presented difficulties, which have now been 

 overcome. The modern all-switch potentiometer is 

 characterized by three features: (1) the resistance 

 is low, yet the switch contacts introduce no error; 

 (2) the thermo-electromotive forces at the switch 

 contact and elsewhere must not cause variations 

 in the reading; (3) the change of setting must 

 not change the resistance of the galvanometer cir- 

 cuit; this in order that the partial deflection 

 method may be employed. These three character- 

 istics can not only be readily secured, but can be 

 obtained in a number of different ways, so that 

 three different types of instrument possessing 

 them are now possible. 



The main point of difference practically is con- 

 cerned with the question whether certain switch 

 contacts shall come in the battery circuit or in the 

 galvanometer circuit. If they are in the battery 

 circuit, much more care must be given to keep 

 their resistance low, but this arrangement is best 

 for reducing the thermoeleetro forces. With the 

 contacts in the galvanometer circuit, their resist- 

 ance is unimportant, and the instrument therefore 



requires less care and attention; the thermo- 

 electric forces can be practically avoided by 

 proper switch construction, so that this arrange- 

 ment seems preferable in a majority of cases. 



(The abstracts of the second and last paper are 

 by their authors.) 



E. L. Faris, 

 Secretary 



THE BIOLOGICAL SOCIETY OF 'WASHINGTON" 



The 383d regular meeting was held April 1 in 

 the lecture hall of the Cosmos Club with President 

 David White in the chair and about a hundred 

 persons present. 



Under the heading brief notes and exhibition 

 of specimens, F. V. Coville showed a pot of trail- 

 ing arbutus {Epigcea repens) in full bloom. The 

 plant had been grown in the greenhouse from the 

 seed, and ooth foliage and blossoms were superior 

 to those commonly found wild. 



The following communications were presented: 



A Day in the Galapagos Islands: William Edwin 



SArroKD.' 



The arboreous cereus of Charles Island has 

 never been adequately described. It was said by 

 Darwin to resemble Cereus peruvianus and com- 

 pared by Engelmann with Cereus multiangularis. 

 It is quite distinct from both these species. The 

 well-known C. peruvianus has only six to eight 

 longitudinal ribs, while the Charles Island Cereus 

 has sixteen to eighteen ribs. It is doubtful 

 whether C. multiangularis is ever arboreous. 

 Weber has described the Charles Island cereus 

 under two names: Cereus thouarsii, which he 

 characterizes as a columnar cereus with superim- 

 posed joints and pleasantly acidulous purplish-red, 

 plum-like fruit containing soft white pulp and 

 numerous small black seeds; Cereus galapagensis, 

 as a plant resembling the arboreous cerei of South 

 America, "with elevated angular stems exceeding 

 the surrounding vegetation. ' ' Both of these de- 

 scriptions apply well to the Charles Island cereus, 

 illustrations of which, in flower, were presented by 

 the author of the paper. The descriptions of both 

 supposed species were published on the same page 

 by Weber, and though C. galapagensis may be a 

 preferable name, as it indicates the habitat of the 

 species, C. thouarsii precedes it on the page, and 

 must therefore be accepted. 



Cereus thouarsii is remarkable for the long 

 slender tube of its funnel-shaped perianth. The 



' This paper is to be published in The National 

 Geographical MagaHne. 



