362 



NA TURE 



{Feb. 1 6, 1888 



should be presented in the form y\a = cosh x^a, using the 

 notation of the hyperbolic functions ; which might also be 

 employed with advantage in the statement of the results 

 of the examples on p. 302. Chains of 5000 feet span, 

 and 400 feet versed sine, are in existence, providing 

 striking numerical examples in this part of the subject. 



Most of the examples are carefully chosen, but the 

 author by diligent search could easily add more interest 

 to the collection, particularly to the examples on para- 

 bolic trajectories, and problems concerning the motion of 

 railway-trains. Ex. 85, p. 499, certainly requires careful 

 revision. The diagrams of the simple machines are of 

 the usual academic nature ; the author should consult 

 Prof Kennedy's " Mechanics of Machinery " for better 

 illustrations, especially of the differential pulley, and of 

 pulley tackle in general. If the differential screw is given 

 (p. 43S), why not also the integral screw, which is to be 

 met with more commonly in real life — for instance, in 

 railway couplings, and in the rigging of ships. 



Except for the parts criticized above, on the units of 

 weight, mass, and force, the present treatise shows that 

 the author 1ms read with profit and discrimination the 

 most recent treatises on dynamics ; he has produced 

 a very useful work, suitable for instruction in technical 

 colleges, and likely also to prove a necessary corrective 

 to the very abstract treatment of the subject of mechanics 

 too common in the character of University instruction. 



A. G. Greenhill. 



ATLAS OF THE DISTRIBUTION OF PLANTS. 

 Atlas der PJlanzenverbreitung. (Berghaus's " Physikal- 



ischer Atlas," Abtheilung V.) Bearbeitet von Dr. Oscar 



Drude. (Gotha : Justus Perthes, 1887.) 



THE history of the science of the distribution of plants 

 begins with Linneeus, who was the first to cite 

 systematically the countries and situations in which the 

 plants he described grew. This we find carefully done 

 in the first edition of the " Species Plantarum," published 

 in 1753. No perceptible advance beyond this was made 

 before the appearance of Humboldt and Bonpland's 

 " Essai sur la Gdographie des Plantes " in 1805, which 

 work may be designated the real foundation of the 

 science. It was followed in 1823-24 by the Dane, 

 Schouvv's " Grundtrsek " and " Plantegeographisk Atlas," 

 the latter containing twenty-two maps illustrating the 

 vegetation of the world, and especially the distribution of 

 plants cultivated for food. There is also a German edi- 

 tion of both the " Outhnes" and the "Atlas." From this 

 date onward many of the most eminent botanists investi- 

 gated distribution in connection with classification of 

 plants, notably R. Brown, A. P. De Candolle, H. C. 

 Watson, C. Darwin, A. De Candolle, J. D. Hooker, 

 Edward Forbes, Von Martius, and Grisebach, to say 

 nothing of the younger botanists. But the results of 

 their labours ars still scattered, or at least only partially 

 elaborated ; for Grisebach, in his " Vegetation der Erde," 

 deals with the facts from a peculiarly narrow stand-point. 

 It is true that both Drude and Engler (" Versuch einer 

 Entwicklungsgeschichte der Florengebiete ") have at- 

 tempted something beyond this, but neither, we suspect, 

 regards his work as more than a preliminary effort. The 

 primary geographical divisions of these two writers are 

 essentially the same, though their nomenclature differs ; 



but, considering the complexity of the subject, probably 

 no two persons would agree exactly on these points ; yet 

 it is highly desirable that there should be something 

 approaching uniformity in the names of the divisions. 

 Grisebach designates his primary divisions " Gebiete," 

 and Drude his " Reiche " ; whilst Engler's four primary 

 divisions are designated " Reiche," and his secondary 

 ones " Gebiete." Let us now briefly examine the main 

 features of Drude's Atlas. Following the most authori- 

 tative English writers on zoological and botanical geo- 

 graphy, we will call the primary divisions regions, and 

 the secondary divisions sub-regions. 



Drude divides the world into fourteen floral regions, 

 and each of these into a number of sub-regions, indicating 

 by lines and dots the overlapping of the elements of con- 

 tiguous sub-regions. The regions are : (i) Northern, (2) 

 Central Asia, (3) Mediterranean, (4) East Asia, (5) Middle 

 North America, (6) Tropical Africa, (7) East African Is- 

 lands, (8) Indian, (9) Tropical America, (10) Cape, (11) 

 Australia, (12) New Zealand, (13) Andes, (14) Antarctic. 



While agreeing in the main with the foregoing divisions, 

 we cannot but regard some of them as including too, 

 much or too little, according to the number of primary 

 divisions adopted. We recognize the difficulties of the 

 task, and admit that it is practically impossible to divide 

 the vegetation of the world into regions of equal value 

 and importance, even leaving out of consideration the 

 mountain flora within the tropics. Instead, however, of 

 giving Madagascar and the neighbouring islands the rank 

 of an independent region, we should treat it as a sub-region 

 of the tropical African flora. On the other hand, the Indian 

 region seems too comprehensive, as it includes the whole 

 of tropical India, Malaya, Cochin-China, the Malayan 

 Archipelago, New Guinea,.North Australia, and Polynesia, 

 even to the Sandwich Islands. The very extensive recent 

 collections of Madagascar plants, made by various English 

 and French travellers, prove that the flora is really a 

 sub-region of the tropical African flora. With regard to 

 the flora of Polynesia, it is true that the littoral element 

 consists almost exclusively of species common to the 

 Malayan Archipelago and North Australia, many having 

 an even wider range ; but the Australian and American 

 affinities of the endemic element are certainly too 

 pronounced, in our opinion, to treat this flora as a sub- 

 region of the Indian ; and the Sandwich Island flora is 

 as highly specialized, to say the least, as that of New 

 Zealand. Perhaps it would be more convenient to 

 make it an independent region. Again, the purely 

 Australian types surely predominate largely over the 

 Asiatic in North Australia, especially if we eliminate the 

 widely-dispersed coast plants. Dr. Drude's New Zealand 

 region includes the surrounding islands, except the more 

 southern Macquarie ; yet, of the eighteen vascular plants 

 recorded from this island, sixteen are common to the 

 New Zealand group. The Auckland and Campbell 

 groups should be reckoned in the Antarctic region rather 

 than New Zealand ; and St. Paul and Amsterdam Islands, 

 as well as the Tristan d'Acunha group, do not belong to 

 the same category. Further, the higher mountain flora 

 of Central America and South Mexico has certainly a 

 greater claim to be included in the Andine region than 

 has that of the Galapagos, though Dr. Drude separates them. 

 We have called attention to these defects or incon- 



