28 



NATURE 



[May II, 1893 



and, owing to the total absence of terrestrial enemies and the 

 abundance of food, may have developed into the allied flightless 

 birds vfhose remains are found there. 



But Mr. Forbes speaks of the genus Aphanapteryx itself, pre- 

 sumably therefore flightless, inhabiting the Antarctic continent, 

 and migrating northwards by two routes of about 2000 miles 

 each in which case, this enormous extent of land must have 

 been as free from all carnivorous land mammals and reptiles 

 as New Zealand and Mauritius are now. If however, the birds 

 in question lost their powers of flight in or near the islands 

 where their remains are found, all difficulties of this kind dis- 

 appear. The Aphanapteryx belongs to a family, the Rallidse 

 or rails, of world-wide distribution, while many of the component 

 genera are also almost cosmopolitan, and are represented by 

 closely allied species in distant regions. What difficulty, there- 

 fore, is there in the same or closely allied species of this 

 widespread group finding their way at some remote epoch to 

 Mauritius and the Chatham Islands, and, from similar causes in 

 both islands, losing their power of flight while retaining their 

 general similarity of structure ? To put the matter briefly : if 

 the common ancestors of the Aphanapteryx of Mauritius and the 

 Chatham Islands were flightless, they could not have reached 

 those islands from the Antarctic continent owing to the length 

 of route and the presence of enemies ; while if they possessed 

 the power of flight no important change in land-distribution is 

 required. 



I have discussed this one point only, because it illustrates the 

 very common practice of explaining each fresh anomaly of 

 distribution by enormous changes of physical geography, when 

 a much more satisfactory explanation can be given involving no 

 such vast and unsupported revolutions in the earth's surface. I 

 am aware that Mr. Forbes adduces many other facts and con- 

 siderations in support of his view as to the former extension and 

 habitability of the Antartic continent, some of which appear to 

 me to be valid and others the reverse. On most of these I 

 have already expressed an opinion in my "Island Life" ; and I 

 only write now in order to point out that the very remarkable 

 and interesting facts, whose discovery we owe to Mr. Forbes's 

 energy and perseverance, do not add anything to the evidence 

 already adduced for that view, but may be best explained in a 

 far simpler manner, and without requiring any important 

 changes in the geography of the southern hemisphere. 



.A.LFRED R. Wallace. 



Swarms of Amphipods. 



Once last winter on entering the laboratory here after it had 

 been shut up for a few days, we found the floor, tables, shelves, 

 window-ledges, and even dishes on the highest shelves, covered 

 with great numbers of dead amphipods. These were found to 

 be Orchcstia gammarellus (the shore-hopper). About ten days 

 ago an unusually high tide occurred, and the curator and others 

 who were working in the biological station noticed that the 

 steps leading to the beach were swarming with amphipods. On 

 investigating further it was found that the amphipods were 

 coming up in great numbers from high-water mark, that they 

 jumped up the steps, and that they climbed the vertical con- 

 crete wall surrounding the station to a height of several feet. 

 Many of them were found about twelve feet above the sea, 

 having come nearly all the way on artificial ground (concrete 

 steps and wall), and they were so abundant on the platform 

 outside the laboratory door that it was impossible to put a foot 

 down without treading on many. Specimens were kept, and 

 Mr. A. O. Walker, who is here with me now, finds that these 

 also are Orchestia gammarellus. This species lives normally at 

 or about high-water mark, and it is abundant here under stones 

 at that line, but Mr. Walker tells me that he has taken it on 

 the one hand nearly at low-water mark, and on the other hand 

 under stones on grass, along with beetles, and we have found 

 it near here far above high-water mark at the side of the road. 

 However, these last are probably exceptional cases, and we are 

 both inclined to think that the two amphipod invasions noticed 

 here have been caused by the Orchestias being driven from their 

 usual haunts by exceptionally high tides. But whether a panic 

 arises on the flooding of their homes, or they lose their way on 

 our concrete, the fact remains that whereas the sea was only a 

 couple of feet higher than an ordinary high tide the amphipods 

 ascended on the one occasion to about twelve and on the other 

 to perhaps twenty feet above their usual level. 



Port Erin, April 29. W. A. Herdman. 



A Difficulty in Weismannism Resolved. 



Weismann's essay " On the Significance of Sexual Repr 

 duction in the Theory of Natural Selection," published in i88 

 enunciates the thesis that the object of sexual reproduction is " 

 create those individual differences which form the material out 

 which natural selection produces new species." This thesis w 

 developed in the essay, " On the Number of Polar Bodies ai 

 Their Significance in Heredity" (1887), and still further 

 "Amphimixis," published late in 1891. 



While " Amphimixis " must have been nearly ready, I wro 

 to Nature (vol. xliv. p. 613), under the heading, " 

 Difficulty in Weismannism," pointing out a posteriori the coi 

 plete insufficiency of sexual reproduction, by merely shufflii 

 ancestral germ-plasms, to effect indefinite specific variation < 

 the lines adopted by Weismann. My friend, Mr. Poulton, wro 

 (vol. xlv. p. 52) accepting my summary of Weismann's vie\ 

 " as fair statements," but criticising the deductions is not alio' 

 ing for the effect of different groupings of the ancestral plasms 

 the germ-cells, and regretting that I had not awaited the pn 

 lication of " Amphimixis," as " Prof. Weismann tells me, i 

 wrote, " that the points raised by Prof. Hartog are consider 

 in this treatise." Mr. Trow also wrote (vol. xlv. p. 102), urgii 

 that I had not allowed for the simultaneous action of natui 

 selection or for the combinations of germ-plasms. In reply 

 my rejoinder of the same date, Mr. Trow again urged that I h; 

 not taken natural selection into account, and that I had m 

 understood Weismann's position. The controversy was th 

 closed. 



However, neither the German edition of " Amphimixis," n 

 the authorised English translation published about s;x mont 

 later, contained the solution of my difficulty that was anticipat 

 by Mr. Poulton. There runs through the book like a red thre 

 the conception of 1886, that sexual reproduction is the creat 

 of the variations on which natural selection acts. A referen 

 of mine to the inadequacy of this, Weismann's Theory 

 Variation, contained in an article in the Contemporary Revi, 

 for July, 1892 (" Problems of Reproduction"), passed withe 

 answer or comment, so far as I know. 



In "The Germ Plasm, a Theory of Heredity" (189, 

 Weismann devotes chapter xiv. to the consideration of heredil 

 Herein I find the following theses, in which I preserve t 

 italics of the original (English edition) : — 



1. "It [sc. amphimixis] is not the primary cause of he 

 ditary variation," p. 414. 



2. " The cause of hereditaty variation must lie deeper th 

 this [amphimixis]. // must be due to the direct effect of extert 

 influences on the biophors and determinants " [sc. of the gei 

 plasms or ids], p. 415. 



3. " The origin of a variation is equally independent 

 selection and amphimixis, and is due to the constant occurrer 

 of slight inequalities of nutrition in the germ plasm," p. 431. 



Obviously the position of 1886-91 has been abandoned 

 untenable. If we ask why, the answer is probably contain 

 in the following passage and annexed note (" Germ Plasir 



PP- 434-5) :— 



" It has recently been maintained that as a consequence 

 my theory I must adopt one of two alternatives, and assui 

 either that the germ plasm of the higher animals consists of i 

 of the primitive protozoan ancestors, or that every id is ct 

 structed in accordance with the existing character of t 

 species ; my real view, however, is intermediate between th( 

 two." The note runs : " Compare Marcus Hartog, Natuf 

 vol. xliv. p. 102." The reference omits my letter of Octol 

 31, 1 89 1. The deductions made by this author from my forn 

 views are logically correct, but are no longer justifiable, sine 

 myself have gained further insight into the problems c< 

 cemed. 



It follows from the above — 



1. That Weismann has withdrawn his whole theory of spec; 

 variation as created by sexual reproduction. 



2. That my account of his views on the point at issue in ifl 

 was both full and fair. 



3. That in 1891 no one else, not even Prof. Weismann, 1 

 perceived that ' ' logically correct " deductions from his gene 

 theory of the germ plasm were fatal to his theory of variatio 



4. That the Weismannism of to-day regards the action 

 external forces as the one essential cause of variation, so 

 approximates to the teachings of the older evolutionists. 



As no reference is made in the preface to this matter, 

 even in the index (for which Prof. Weismann is not responsih 



NO. 1228, VOL. 48] 



