12 



NA TURE 



{May 5, 1887 



results he found were 527 and 51-5. The mean initial 

 temperature in these two experiments is - 8°7 C, and the 

 mean final temperature -7°"2 C. At -7°'2 C., ice would 

 form on cooling, and would melt on warming a solution 

 of chloride of sodium containing 6-48 per cent. CI, which 

 represents 11 '87 per cent, of the sea salt. In order to 

 concentrate a brine containg 3 53 per cent, salt to one 

 containing ir87 per cent., ^o per cent, of the water in it 

 must be removed. Hence in sea-water freezing at a 

 final temperature of -7°"2 C, there is formed 70 per cent. 

 of ice, and there remains hquid 30 per cent, of brine. 

 Freezing began at the mean temperature - 8°7 C, and the 

 latent heat of pure ice at this temperature is 75. Calcu- 

 lating the latent heat of this mixture from the heat liberated 

 in the calorimeter during freezing, and assuming that the 

 whole mass had solidified, Pettersson's results give the 

 mean latent heat of this sea-water as 52' i. Calculating 

 the apparent latent heat on the assumption that 70 per 

 cent, of the mass solidifies into pure ice and that 30 per 

 cent, remains liquid, we get the number 51-5. On all 

 grounds therefore we must conclude that pure ice is the 

 primary product in freezing sea-water and saline solutions 

 of moderate concentration. 



The fact that ice melts in sea-water at a temperature 

 of — i°"6 C. to — i"'8 C, according to its saltness, 

 explains the anomalous distribution of temperature in 

 Antarctic waters, and furnishes an account of the motive 

 power which draws the surface water of cold temperate 

 regions into the deeper layers, and after dilution with the 

 melted matter of the icebergs, to the surface layers of 

 Antarctic latitudes. Forming as it does an important 

 factor of oceanic circulation, this part of the subject was 

 treated in a separate paper, of which an account has 

 already been given in Nature (vol. xxxv. p. 516). 



THE CLASSIFICATION OF SPIDERS.^ 

 A LTHOUGH Dr. ThorelFs paper is nominally only a 

 -^-*- criticism of Dr. Bertkau's views, it is really a masterly 

 sketch of the literature on the subject of the classification 

 of spiders, and a review of the methods of various authors 

 from Lister downwards. The two leading views at 

 present held on this subject appear to be represented by 

 the author (Dr. Thorell) and Dr. Bertkau. These two 

 views may be stated generally as atiatomical v. biological, 

 the former being the basis of Dr. Bertkau's classification, 

 the latter (combined with considerations of external struc- 

 ture) that of Dr. Thorell. Dr. Thorell successfully, as it 

 appears to us, defends the classification provisionally 

 adopted in 1869, in his work "On European Spiders,'' 

 from the sweeping charge that it is neither natural nor 

 equal, nor based on characters of sufficient importance 

 and distinctly expressed, though at the same time he 

 freely admits its inevitable imperfection. He shows that 

 in no branch of Nature are the subordinate eroups of 

 exactly equal value, nor should it be expected that the 

 sarne equal systematic value would be found in the sub- 

 ordinate groups of the class Arachnida. It is well to 

 remember that zoologists have to form their groups out of 

 such materials as have come to their hands ready pro- 

 vided for them by Nature. They cannot expect to advance 

 natural science by constructing out of limited materials a 

 perfectly symmetrical system, and then insisting that all 

 the diverse forms of Nature shall, nolens volens, be stufi"ed 

 somewhere or other into it. Equality, therefore, of sys- 

 tematic value in the various groups into which spiders 

 (Araneidea) may be divided can at best be only approxi- 



T,' ^'£" ^r; Bertkau's Classification of the Order Aranea;, or .Spiders," by 

 ?hl'L7 •'■^ ^^rV'"''-^^^^-,^-^^ \^P."' ^886, pp. 301-26), (See especially 

 the following works by Dr. Bertkau :-" Versuch einer naturlichen Anordnung 

 aer bpinnen, Archw. fiir Natur^eschichte, xliv. i, pp. 351 et scq 1878- 

 and Ueber das Cribellum und Calamistrum ; ein Beitrag zur Hi.stoIo<'ie,' 

 fse^T^' Systematik der Spinnen," ibid., xlviii. i, pp. ^i^etseq\ 



mate ; and it seems evident that as our knowledge of 

 structure, whether external or internal, or of habits as 

 dependent on and arising out of structure, becomes more 

 extensive and exact, so some further modifications may 

 become necessary in the primary divisions of spiders. 

 After subjecting Dr. Bertkau's classification (which is 

 based principally on the breathing-organs) to a minute 

 and destructive criticism. Dr. Thorell modifies his former 

 views by reverting, in some measure, to the Latreillian 

 division of spiders into (i) those possessing four air-sacs, 

 Tetraptieunwnes J (2) those with two, Dipneuniones j still, 

 however, retaining the old Latreillian biological divisions. 

 TerritelaricE , Tubitelaria:, OrbitelaricF, &c., based on 

 habits, because the groups so divided may yet be 

 thoroughly and scientifically differentiated by important 

 and trustworthy structural characters. These divisions 

 (now called by Dr. Thorell " tribus ") are, as is well known, 

 seven in number, and each is subdivided into families, 

 which, with few exceptions, include only European 

 species. Dr. Thorell's opportunities for the study of exotic 

 groups not enabling him to construct a more complete 

 subdivision of all known spiders. Dr. Bertkau's primary 

 division of the Araneidea is into two groups, called sub- 

 orders — Tetrasticta, with four breathing-apertures, and 

 Tristicta, with three. Dr. Thorell shows conclusively 

 that in the present state of our knowledge of the respir- 

 atory system of spiders (though this is far advanced 

 beyond what it was in Latreille's days, and in a great 

 degree the advancement is admitted to be due to Dr. 

 Bertkau's labours) these two suborders are artificial 

 rather than natural ; as are also his subdivisions of the 

 Tristicta, which are based on the undoubtedly remarkable 

 characters to which such great prominence was given by 

 the late Mr. Blackwall, that is, the possession or absence 

 of a cribellum and calamistrum, the use of these in 

 primary subdivision bringing together spiders not closely 

 allied by any other natural characters. Dr. Thorell 

 attributes a certain amount of classificatory importance to 

 these organs by his intercalation of the families of his 

 largest " tribus " possessing them, in a kind of osculant or 

 collateral way, and of the other '• tribus " in which they are 

 found, in a linear arrangement, guided, however, in both 

 cases by their possessing such other characters as, in all 

 instances, fully warrant the position assigned to them. 

 The anatomical study of the trachecB (properly so called) 

 of spiders seems to be yet in its infancy. Certainly 

 at present these organs of respiration do not appear to 

 warrant the importance attributed to them by Dr. 

 Bertkau ; and although Dr. Thorell's primary subdivisions 

 are, in their naines, strictly speaking, based on only bio- 

 logical characters, yet in reality they severally enshroud 

 the most important structural ones, and hold all known 

 spiders in a fairly natural system. They are, moreover, 

 well known, and appear likely to be adopted for some 

 time yet to come, with more or less modification, by the 

 majority of araneologists. 



O. P. Cambridge. 



CHRISTMAS ISLAND. 



HTHE following account of the little-known Christmas 

 *■ Island, situated in the Indian Ocean, south of Java, 

 may be of interest. Capt. Maclear and his officers col- 

 lected a variety of specimens, which have been forwarded 

 to the Museum of Natural History and to the Royal 

 Gardens, Kew, but they do not seem' to have succeeded 

 in making their way through the dense jungle to the 

 upper part of the island, to ascertain the geological cha- 

 racter of the mountain originally protruded from the 

 depths. It is a little remarkable that, in a sea so calcu- 

 lated to encourage coral growth, no new reefs should 

 have formed round the island since the ancient ones 

 were elevated above the surface The Cocos or Keeling 



