492 



NA TURE 



[June 24, 1909 



where K ij nie radius of the tube, L its length, and /i 

 and p' the pressures at the ends. By a series of experi- 

 ments with a tube, the length of which was 2-oo cm., 

 the radius 000333 '^"i-> 'he proportions between the 

 measured values and those calculated from this formula 

 were found to be, for hydrogen, 104 ; oxygen, loi ; 

 carbonic acid, loi. 



The formula will, however, only apply correctly when 

 the radius of the tube is small compared with the mean 



p' + p" 

 free path. With increasing pressure - - ■ (decreasing 



mean free path), the gas flow of a given value for p' — p' 

 decreases to a minimum, and afterwards increases in order 

 to approach the value which it should hqve according to 

 Poiseuille's well-known law. That this must be the case 

 may easily be inferred from the kinetic gas theory in con- 

 nection with the above-mentioned theory as to the inter- 

 action between gas molecules and a wall. 



Martin Knudsen. 

 The University, Copenhagen. 



• The Germ-layer Theory. 



The germ-layer tlv-ory as stated on p. 428 of Naturb 

 (June 10) by Mr. Stanley Gardiner appears in a rather 

 extreme form. Probably all will agree that, not only the 

 germ-layer theory, but every theory of development, pre- 

 supposes a certain definiteness in structure of germ cells. 

 But if that much is granted, it is not necessary to sup- 

 pose that the differentiation of protoplasm has proceeded 

 at so early a stage to such an extent as to preclude abso- 

 lutely the possibility of protoplasm, which has been so 

 far misplaced by experiment as to find itself in a new 

 environment, responding to the influences of the new 

 environment and so developing along a path it would 

 not have followed had the experiment not been performed. 



It seems difficult to comprehend what reason can be 

 assigned for regarding those organs of the early phase of 

 the life-cycle which we call germinal layers as being less 

 capable of showing homology than the organs of later 

 phases which we speak of as adult. 



Surely the biological principles — whatever they may be 

 — must apply equally throughout all periods of the life- 

 cycle. 



The argument from regeneration is hardly conclusive, 

 because one essential of regeneration and budding seems to 

 be the regression of differentiated protoplasm into un- 

 differentiated protoplasm (or, at any rate, the origin in 

 some way or other of an undifferentiated cell mass), that 

 is to say, a regression to a state equivalent to a segment- 

 ing egg, namely, a state really prior to that of germ- 

 layer formation. 



Finally, it must be remembered that visible differences 

 and resemblances are much less obvious in these early 

 phases of the life-cycle than later, and that the difficulty 

 of observationi owing to the minute size of the objects, is 

 so great that errors of observation, which delay correct 

 interpretations, are far more frequent than is the case 

 with work upon the grosser phases of the life-cycle. 



It cannot be conceded that the " anomalies in the forma- 

 tion of the layers in vertebrates " which are " patent to 

 every student " are all capable of substantiation. 



Grantchester, June 15. Ric. Asshetox. 



The Pollination of the Primrose. 



It appears that in a previous note on this subject 

 (N'atl-re, June 17, p. 457) clearness may have been sacri- 

 ficed to brevity. It is not meant that humming-bird and 

 bee hawk-moths can be regarded as usual or frequent 

 agents in the pollination of the primrose. They arc men- 

 tioned in proof that some moths do, now and then, visit 

 the flowers, and may presumably aid in their cross- 

 pollination. There can be little doubt, however, that the 

 humble-bee is herein the chief agent, and in this district, 

 I should say, more particularly Bombus horlorum. 



\V. E. Hart. 



KiMerry. Londonderry, June 19. 



NO. 2069, VOL. 80] 



nWST A.XD ICE CRYSTALS.' 

 T"\l RING (he past quarter of a century Mr. Wilson 

 ^-^ J. Benlky h;is devoted himself with a patient 

 industry deservini,' of all praise to securing permanent 

 records of the multitudinous forms assumed by water 

 in its crystallised condition. The work has been 

 executed at his home, a farmhouse, situated sixteen 

 miles east-north-tasi of Burlington, Vermont, near 

 the Canadian border, at an altitude of 1500 feet above 

 sea-level, where the low temperatures experienced 

 every winter are very favourable for the study of these 

 forms. Seven years ago we directed attention 

 (Nature, 1902, vol. l.w., pp. 264-6) to his beautiful 

 series of photomicrographs of snow crystals ; a selec- 

 tion of them w.is reproduced in the U.S. Monthly 

 Weatlicr Iivvicii.\ and was accompanied by a paper 

 in which Mr. Bentley described the methods used for 

 obtaining the photographs, and the facts that could 

 be established from a study of the almost bewildering 

 variety of the forms represented. At the same time, 

 but mainly durini; the subsequent years, Mr. Bentley 

 has been further engaged in preparing a companion 

 and complementary series of frost and ice crystals, 

 i.e. the forms assumed by water that has crystallised 

 immediately upon the surface of the earth. A large 

 number of different types were reproduced in 

 successive numbers of the Monthly Weatlier Review 

 from August to December, 1907, and Mr. Bentley 

 again contributes a description of the apparatus used, 

 and full details with regard to the circumstances under 

 which the several pictures were obtained. 



Nearly the wlwie of the present series represents 

 crystals that were formed during the winters of 1904-5, 

 1905-6, and 1906-7. For several reasons fewer diflicul- 

 ties were experienced in obtaining photographs of 

 these crystals than was the case in the investigation of 

 snow crystals ; they could invariably be photographed 

 in the positions in which they were found, and since, 

 owing to the gre.uer duration of growth, their size is 

 usually much lart^er, smaller magnifications were re- 

 quired, and, indeed, in pictures of groups of crystals 

 actual reductions were called for. The apparatus 

 used was consequently simpler in character. For the 

 majority of the photographs, in which the magnifica- 

 tion did not exceed eight diameters, an ordinary por- 

 trait-lens was used in a camera which was fitted with 

 a homo-made extension arrangement, and the crystals 

 were illuminated obliquely. For higher magnifications 

 a microscope-objective, of J- or i-inch focal length, was 

 employed, and the illumination was direct. The 

 second method, which was required for the minute 

 flakes deposited on windows, entailed more trouble in 

 manipulation, because, while the camera was indoors, 

 the diaphragm fur cutting off all but direct light was 

 on the other side of the window, and had, of course, 

 to be adjusted for each position of the camera. 



The series is divided into three principal groups — 

 hoar-frost, window-crystallisation, and ice — a few sec- 

 tions dealing with li.iil being appended, and for con- 

 venience each group is split up into divisions and 

 subdivisions, according to the shape or the grouping 

 of the crystals. The hoar-frost group is divided into 

 two main divisions -tabular and columnar — but the 

 distinction is appartnily one of degree only, and can- 

 not be pressed. We have selected as an illustration 

 of this group a beautiful example of the " open 

 branch or tree-liki " structure (Fig. i). It will be 

 noticed that the stiiiis broaden out into well-developed 

 plates at their terminations. The study of the crystals 

 deposited on windows obviously admits of greater ease 

 of observation, and, since the conditions of the atmo- 



1 " Studies of Frosl and lie Crystals, 

 nth 373 figures on 31 ]' 

 terim; 1907.^ 



Hy Wilson J. Renllcy. Pp. 11 : 

 (Reprinted from the Monthty Weather 



