July 21, 192 1] 



NATURE 



651 



micrometer with lines 6ojU apart be used over a 2-mm. 

 objective with No. 8 eyepiece, each division represents 

 o-7;U, and a fifth part of a division corresponds to only 

 I /175,000th of an inch on the slide, which is there- 

 fore at the tube's centre the extent of the necessary 

 distortion for bracket and bearing. The original posi- 

 tion is recovered completely when pressure is removed. 

 Probably evervone who uses high powers of wide 

 angle has acquired the habit of effecting similarly 

 extremely fine adjustments of focus by pressure on 

 the stage. 



With a lower power it is often helpful to press 

 slightly on the nosepiece instead of moving the slide. 

 For quick, rough measurements in the course of 

 other work Prof. Dixon's "ghost micrometer" 

 is very valuable (my friend Dr. W. R. G. Atkins 

 introduced me to it). When the light is taken from 

 a window the image of a piece of wire-gauze leaning 

 against the pane can be brought on the object hy a 

 turn of the mirror, and removed again without losing 

 sight of the object. 



It may be worth adding that in measuring a dis- 

 tance in the line of sight (thickness) by the scale on 

 the fine adjustment, one notch of the milling on the 

 fine-adjustment head corresponds in my Zeiss to 

 i/ioth of a division, or to i/x (i-5/i with a dry objec- 

 tive). The notches can be read opposite the pointer 

 through a lens, with a probable total error of o-4/x 

 for the single measurement of thickness. 



In measuring the width or thickness of a calcite 

 spicule the optic axis of which is parallel to the plane 

 of the slide, or in examining an object above or below 

 the spicule, greater accuracy can be obtained by plac- 

 ing on the ocular a Nicol wjth the plane of polarisa- 

 tion at right angles to the optic axis of the spicule, so 

 that the high refraction of the ordinary ray is 

 abolished. Ebner (5. JB. Ah. Wiss., Vienna, vol. xcv., 

 p. 73) recommended to spongologists the use of the 

 single Nicol for determining the direction of the optic 

 axis of spicules. 



Inierference-colours between Nicols. — The measure- 

 ment of thickness in these sponge-spicules is very 

 difficult, and I hope to substitute for it the 

 mere reading of the spicule's colour between Nicols. 

 Empirically, the colour of all but the two "limbs" 

 of the cylinder appears to be closely that of a 

 calcite plate of equal thickness, and to be irrespec- 

 tive of the angular aperture of the objective (020 to 

 I 40) and of the presence or absence of an Abbe con- 

 denser between the polariser and the object. These 

 are, to me, unexpected results in view of the much 

 longer path in the spicule taken by the ordinary ray 

 as compared with the almost unrefracted extra- 

 ordinary ray. As it is difficult to exclude a 5 per 

 cent, error from determinations of either the thickness 

 or the retardation, and as it has been disputed 

 whether the carbonate of lime in a spicule be wholly 

 calcite, I shall be very grateful if a physicist will 

 supply the theory of the colour of a cylinder (elliptical 

 or circular) of calcite in Canada balsam. The dia- 

 meter of the cylinder ranges upwards from a wave- 

 length of light to lo/i or so; the lowest _A I have 

 determined accurately is i34±4/i/x for the middle band 

 of a spicule, which is a right cylinder since its optic 

 axis is parallel to its length, and the width of which 

 is 830±40yu//. Geo. P. Bidder. 



Cambridge, July 10. 



Ocean Tides. 



In the letter in Nature of May 26 (p. 393) under the 

 above heading, by Mr. H. A. Marmer, of the U.S. 

 Coast and Geodetic Survey, it is pointed out that tidal 

 observations would be greatly enhanced in value^ if 

 permanent bench-marks were established in connection 



NO. 2699, VOL. 107] 



with them, not only for the correlation of any future 

 tidal observations at the same places, but also for the 

 determination of the rate of elevation or subsidence of 

 the land relatively to the sea. 



It may be of interest to note that this question was 

 taken up some fifteen years ago by the Academy of 

 Sciences of Paris, when a prize was offered for the 

 best determinations of mean sea-level from tidal ob- 

 servations in any country bordering on the North 

 Atlantic as a basis for such relative change in eleva- 

 tion on its coast-line. This prize was awarded to the 

 present writer as superintendent of the Survey of 

 Tides and Currents in Canada, as it was found that 

 we had already tidal data available for this purpose, 

 because referred to permanent bench-marks, on an 

 extent of eight deg^rees of latitude from southern Nova 

 Scotia to Belle Isle Strait. 



Although this survey was primarily org-anised in the 

 interests of navigation, its f>ractice of establishing' 

 local bench-marks from the outset in 1894 is also 

 bearing- fruit in other directions. It has afforded to 

 the Geodetic Survey of Canada, more recently 

 org^anised, determinations of mean sea-level on both 

 Atlantic and Pacific coasts as a basis ready to hand 

 for extended levelling^ throughout Canada. Our 

 Geological Survey also refers its contoured maps to 

 mean sea-level, and in several regions it has been 

 possible to give that survey an independent starting- 

 point for these contours from tidal observations 

 already obtained at a locality in the region, as they 

 were referred to a local bench-mark. 



As it is not often that the same superintendent 

 remains in charge of a survey for so long as twenty- 

 seven years, it may be allowable to give these examples 

 in our experience of the advantages of the practice 

 recommended bv Mr. Marmer. which mav accrue years 

 afterwards. ' W. Bell Dawson. 



Ottawa, Canada, June 22. 



American and British Superannuation Systems. 



The writer of the leading article on this subject in 

 Nature of June 30 may have misled your readers by 

 the last paragraph but one in his article, because : — 



(i) No money can go into the pockets of share- 

 holders of mutual insurance companies ; there are no 

 shareholders. 



(2) If an endowment assurance is taken under the 

 Federated System the benefits are increased by the 

 share of profits, which, in the case of a mutual com- 

 pany, means a full share of all profits made. 



(3) The expenses of the selected insurance com- 

 panies are probably little more than those necessitated 

 by a separate "association" when we bear in mind 

 that the premiums charged under the Federated 

 System allow for the saving of "commission to 

 agents " by those offices that usually employ agents. 



(4) The Federated System obtains the advantage of 

 the experience of insurance companies in investing 

 money expeditiously on a large scale. 



The objection quoted as having been made by Mr. 

 Fisher on the second reading of the School Teachers 

 (Superannuation) Bill, 1918, to the effect that public 

 money would go in "dividends to the shareholders" 

 is met by (i) and (2) above. The real difficulty of 

 placing his pension arrangements in the hands of 

 insurance companies is that they cannot assess the 

 invalidity risk or the future salaries on which the 

 pensions in the Bill depend. The invalidity risk, the 

 future salary scale, and the longevity of pen- 

 sioners have necessitated far heavier contributions to 

 pension funds than had been expected when the funds 

 were started, and the difficulty of keeping private 

 pension funds in a solvent condition is so well known 



