538 



NA TURE 



[September 29, 1904 



the application to meteorology of the theory of correlation. 

 The two stations taken were Wilmington (North Carolina) 

 and Halifax (Nova Scotia), distant from one another about 

 1000 miles. Allowing different intervals between the corre- 

 sponding barometric observations, the magnitude of the 

 correlation was found to vary with these intervals, being 

 greatest when Halifax is taken one day later than Wilming- 

 ton. It was suggested that for selected places further apart 

 the discovery of correlations of sufficient magnitude might 

 be of use in the practical work of prediction. 



Major B. Baden-Powell described briefly the development 

 of the aeroplane, and gave an account of the experiments 

 that he has been recently carrying on. Chief among these 

 were his gliding experiments made at the Crystal Palace, in 

 which he is seeking to find out how a man-carrying machine 

 behaves while travelling in the air. 



After a paper by Prof. D'Arcy W. Thompson on Plato's 

 theory of the planets, the business of the subsection came 

 to a conclusion, the following papers being taken as read : — 

 Report of Committee on I'nderground Temperatures ; Dr. 

 F. Hirtel, Zur Flugfrage ; Rev. J. M. Bacon, upper air 

 currents and their relation to the audibility of sound ; Prof. 

 I.emstrom (Helsingfors), on the effect of electric air- 

 currents; J. Hopkinson, the rainfall of the midland and 

 eastern counties of England, and the rainfall of England, 

 1861-1900. William J. .S. I.ockvf.r. 



'/.OOLOGY AT THE BRITISH ASSOCIATION. 

 'T'HE meetings of .Section D were held in the new Sedg- 

 wick Museum of Geoloyv. On Thursday morning, 

 August 18, Mr. W. Bateson, F.R.S., delivered to a large 

 audience his presidential address (see Nature, August 25), 

 the vote of thanks for which was moved by Prof. Poulton 

 and seconded by Prof. Max Weber. In the afternoon Prof. 

 F. W. Keeble gave an address, illustrated with diagrams 

 and lantern slides, on the coloration of marine Crustacea, 

 embodying results of the re.searches carried on bv Ur. 

 (lamble and himself during the last seven years. He de- 

 ■scribed the prawn-like Hippolyte -uarians which lives among 

 the seaweeds around our coasts, and matches their colour 

 with marvellous precision. Its colours range through green, 

 yellow, and brown to red. When given a choice between 

 different coloured weeds, this animal invariably picks out 

 for its abiding place that weed which is in harmony with 

 its own coloration, a faculty to be ascribed not to the 

 possession of a colour sense by the animal, but rather to an 

 •extreme sensitiveness to light. Only in one position, 

 namely, on weed of its own colour, is it in a position of 

 light equilibrium, and then it rests. The coloration of the 

 animal is produced by the manipulation of its three colour 

 pigments, red, yellow, and blue. The common shrimp and 

 prawn were shown to possess the same mechanism, although 

 they make little use of it. Transparent young Hippolvtes 

 placed on weed of a certain colour develop the pigment 

 necessary to approximate the animals to that colour in forty- 

 eight hours or less ; older animals tgke almost a week for the 

 same process, and adults a fortnight, but even then the 

 result is imperfect. Though the adults have lost, to a large 

 extent, this power of sympathetic colour change, the pig- 

 ments react rapidly to changes of light. This reaction is 

 most marked not when the intensity of light changes, but 

 when the background on which the animal rests is changed 

 from white to black, i.e. from one which scatters to one 

 which absorbs light. In the daytime the pigments are ex- 

 panded ; at night they are contracted, except the blue, so 

 that the animal has then a transparent azure colour. This 

 is a true periodic change ; it has become a habit, and 

 endures for days even though the animals be kept in dark- 

 ness. 



Prof. W. B. Scott, of Princeton, U.S.A., then delivered 

 an address (with lantern illustrations) on the Miocene 

 ungulates of Patagonia The animals described were 

 collected by the Princeton E.xpedition from the Santa Cruz 

 beds, the Miocene age of which seems to be now established. 

 Prof. Scott pointed out that while these South American 

 ungulates are singularly different (especiallv in the struc- 

 ture of the periotic region) from those of the' northern hemi- 

 sphere, it is not unlikely that they have a common origin, 

 , as .Anipghino has described a number of genera from pre- 

 NO. I 82 2. VOL. 70] 



Patagonian formations which, though incompletely known, 

 appear to be referable to the Condylarthra — the parent stock 

 of the northern ungulates. Very probably an early Eocene 

 or late Mesozoic migration carried the Condylarthra into 

 .South America, and there, in almost complete isolation, they 

 graduallv gave rise to the various peculiar orders of the 

 Noto-Ungulata. The possibility of such a migration is 

 shown by the discovery of an armadillo in the Middle Eocene 

 of North America. 



The section was occupied practically the whole of Friday, 

 August 10. with papers and discussions on heredity, Prof. 

 Hickson, F.R.S., in the chair. 



The first paper was by Miss E. R. Saunders on heredity 

 in stocks. She said that since the re-discovery of Mendel's 

 work, experimental evidence of the purity of the germ cells 

 has been found in a rapidly increasing number of examples. 

 Much of this evidence has been derived from cases like those 

 studied by Mendel where the differentiating characters are 

 related to each other as dominant and recessive. In such 

 cases the individuals of the (F,) first generation (DR) show 

 the dominant character, and those of the second (F,) gener- 

 ation the two parental characters in the ratios 3 D ; i R or 

 I D : I R, according as they result from DRxDR or 

 DRxR. In other cases the results are complicated by 

 reversion, gametic coupling of distinct characters, &c., and 

 they require careful analysis, and several generations may 

 be required to elucidate them. As a surface character 

 hoariness is dominant, glabrousness recessive. Experiments 

 in the form DRxDR or DRxR, where D is the white- 

 flowered form of .Matthiola incaiia and R a glabrous ten- 

 week strain, give normal Mendelian ratios in F,. In other 

 cases the result, as regards hoariness and glabrousness, is 

 more complex, owing to the different behaviour of various 

 glabrous strains, which, as far as can be seen, differ only 

 in flower colour. As to flower colour, various combinations 

 of colours give reversionary purple in the first generation 

 (F,). Purple F, may also be produced by two white parents 

 if thev belong to strains differentiated by the leaf surface. 

 Such purple cross-breeds may give a simple Mendelian result 

 in F., or a varietv of new colour forms may appear, this 

 latter beiner commonly seen when cream is one of the 

 parental colours. For example, in a cross of a glabrous 

 white with a glabrous cream, at least nine colour forms 

 were produced in F... Whether the appearance of these new 

 forms indicates disintegration or simply re-combination of 

 preexisting characters is uncertain. Creams breed pure at 

 once. Some whites are pure, others are heterozygotes with 

 cream. The number of extracted recessive types resulting 

 from a given union and their specific behaviours are not yet 

 known. 



Mr. A. D. Darbishire gave some account of his experi- 

 ments on the breeding of mice. The Japanese waltzing 

 mice show the well known restless and spinning movements ; 

 thev have a piebald vellow and white coat and pink eyes. 

 When an albino is crossed with a Japanese waltzing mouse 

 the majoritv of the offspring are on first inspection in- 

 distinguishable from the common house mouse, and they 

 invariably (in all the 300 cases bred) have black eyes. 

 Hybrids never exhibit waltzing movements, and they are 

 never albinos. When such hvbrids are bred together they 

 produce offspring which, considered from the point of view 

 of colour, fall into three categories: — (i) those (half the 

 number) with black eyes and coloured coat, and therefore 

 resembling their parents; (2) those (one-fourth) with pink 

 eyes and coloured coat, therefore presenting the same 

 features of eye- and coat-colour as Japanese waltzers ; (3) 

 those (one-fourth) with pink eye and uncoloured coat, i.e. 

 albinos. About one-quarter of these hybrids waltz, but the 

 rest are normal in their progression, and the waltzing habit 

 may be associated with any of the three colour categories. 

 The albinos (group three) breed true, the pink-eyed mice 

 with coloured coats breed nearly true, and the biack-eyed 

 mice with coloured coats produce, when paired together, 

 albinos, pink-eyed mice with coloured coats, and black-eyed 

 mice with coloured coats (proportions of each not yet deter- 

 mined). Some of the facts seem to confirm the Mendelian 

 interpretation, while others may be described in terms of 

 either Galton's or Pearson's formulje of ancestral inherit- 

 ance. 



Mr. C. C. Hurst described some experiments on heredity 

 in rabbits. .\n inbred pair of albino Angoras was crossed 



