iqS 



NA TURE 



[June 28, 1894 



are now ventured as ihe result of perusing pp. 42-48, which 

 relate to the discontinuity of certain colour- variations. 



Without attempting (o discuss Mr. Bateson's gener.il pro- 

 positions, I desire to point out that the facts related in the 

 portion of the work cited, and the " many similar cases " which 

 might be added, do not altogether support the idea of discoii- 

 linuoiis progressivt colour-variation, as distinguished from 

 a/afiim. 



Various writers, including myself, have on sundry occasions 

 endeavoured to demonstrate that both in plants and animals a 

 definite succession of colours may be observed. In flowers, 

 for instance, from pink to purple, and from yellow to red ; 

 in birds and insects, and many molluscs, the yellow to red 

 succession is commonly observed. In such instances as these, 

 it has been held that one colour represents a lower stage of 

 evolution nr a less degree of metabolism than the other ; and it 

 has been many times pointed out, that discontinuous iitai'iilic 

 variation, <-./. from red to yellow, is commonly to be seen. 



Now I take it that .Mr. Bateson considers the evidence which 

 he adduces, to illustrate the frequency of discontinuous pro- 

 gressive variation in colour, not merely reversion. Let us 

 examine this evidence a little more closely. 



According to the views held by the writers above mentioned, 

 red is "higher" than yellow, and red varying to yellow is 

 reversion. Such reversion is well known to be often discon- 

 tinuous, as in the yellow-fruited yew, the yellow tomato, the 

 yellow-fruited raspberry, the yellow varieties of various red 

 moths, and so forth. 



But is the yellow to red variation, which is supposed to be 

 the progressive one, discontinuous? L^t Mr. Bateson himself 

 tell us. On p. 45 he cites the variations of the yellow Gonep- 

 teryx rhamni towards orange. .Vre these disCDntinuous ? Do 

 we find among the yellow rhamni some that are entirely 

 orange? Not so, " there are records of specimens . . . more 

 or less flushed with orange. ' 



Exactly : whereas among the red species of Callimorflia, 

 Arctia, Zys^cena, &c., we find varieties not flushed with yellow 

 but entirely yellow in place of red (the dark markings being of 

 course as usual), in the yellow 6'. rhantni we find continuous 

 variation towards orange, none yet having attained actual 

 red. 



In birds red species may vary to yellow ; green also to 

 yellow, and such variation may be sudden. But yellow to 

 green? or yellow to red? We have got our canary yellow 

 easily enough, but all the art of the breeder cannot get him 

 redder than orange, and the variations thereto are fairly con- 

 tinuous. 



We cannot get a blue rose ; but the blue Delphinium, the blue 

 Penlstemon, these readily var)' to pink. We may have a yellow 

 rose, but it is pretty well agreed that if we ever do see a blue 

 one, it will be by a processof((Vi/;'////('//.c variation and selection. 

 Will not Mr. Bateson admit that he would be immensely 

 astonished to see a blue rose arise from seed of a red one, or a 

 scarlet canary from eggs laid by a yellow one? Yet red from 

 blue, or yellow from rc(i, would seem scarcely worth comment 

 in any group of animals or plants, .so numerous are the recorded 

 instances of this kind of variation. 



On p. 44, Mr. Bateson cites instances of blue in place of red, 

 which should be progressive variati m. This occurs in Calocala 

 nupta, for instance, but very rarely, and instances which seem 

 rather intermediate arc on record. Another sample cited 

 \s the blue-flowered Ana^^allis arveasis. Here the case is 

 different, for the blue and red varieties are entirely distinct, 

 and come true by seed. I have myself lived in districts where 

 the blue and red varieties respectively abounded, and in neither 

 locality did I ever sec intermediates. They had all the appear- 

 ance of true species, which they have often, I think with justice, 

 been considered. The locality for the blue variety was Funchal, 

 Madeira, and there the red pimpernel also occurs. But in 

 England, where the red variety is so common, I never saw the 

 blue one truly wild. 



In Primula we have yellow species and red species, and, as 

 everyone knows, our common primrose may vary to red. But 

 alio, as everyone knows, the variation is continuous. How 

 well I remember as a child looking for those that were tinged 

 with red, always hoping to get one redder than that last 

 toond : 



The lubject admits of much greater amplification than is now 



po--''-- -• ' ■• - ' .. . - i.-nicd that many instances 



m.i :tvailable which appear to 



in'l , r^ - i 'iir-v.Tri.-ition. But never- 



theless, taking the evidence as a whole, I will venture to urge 

 the validity of the following statements : — 



(1) Colour-variation occurs in a definite order, the colours 

 forming one or more series. 



(2) Variation from those lower to those higher in the scale of 

 evolution, or from those representing less to those representing 

 greater metabolism, is usually continuous. 



(3) Reversion from a higher to a lower colour is usually dis- 

 continuous. T. D. .-v. COCKERELL. 



Las Cruces, New Mexico, U. S. .\., June i. 



Niagara River since the Ice Age. 



Much new light on the Quaternary history of the great lakes 

 tributary to the St. Lawrence river has been contributed in 

 three recent papers by Mr. V. B. Taylor, all published within 

 the short time since Mr. G. K. Gilbert's writing on " The 

 Niagara River as a Geologic Chronometer,' which appeared in 

 NATtJRE for May 17 (page 53). These papers are in the 

 Bulletin of the Geological Society of America (vol. v. pp. 620- 

 626, Aprii 30, 1894), and in the American Geologist (vol. xiii. 

 pp. 316-327 and 365-383, May and June, 1894). Supplement- 

 ing the earlier observations and studies of Whittlesey, New- 

 berry, Gilbert, Spencer, Lawson, Leveretl, Wright, Baldwin, and 

 the present writer, among others, these latest explorations and 

 discussions by Mr. T-iylor enable us to form a very definite and 

 closely connected historical statement of the relationships of 

 the ice-dammed lakes which preceded the present Laurentian 

 lakes, and of their dependence on the gradual departure of the 

 ice-sheet and on the accompanying northward uplift of that 

 region. 



The largest element of uncertainty in the estimate of 7000 

 years for the Post-glacial period, from the retreat of the ice- 

 sheet to the present time, drawn from the rate of recession of 

 the Falls of Ni.igara, consists, as Mr. Gilbert has shown, in the 

 probability or possibility that for some considerable time, next 

 lollowing the melting away of the ice upon the area crossed by 

 the Niagara river, the outlet of lakes Superior, Michigan, and 

 Huron may have passed to the St. Lawrence by a more northern 

 course, flowing across the present watershed ea^t of lake 

 Nipissing to the Mattawa and Ottawa rivers. Mr. Taylor's ob- 

 servations now indicate, however, if interpreted on the 

 hypothesis of gkacial lakes (which is believed by Mr. Gilbert and 

 by the majority of other geologists of America to be the true 

 view), that the glacial lake Warren, filling the basins of 

 Superior, Michigan, Huron, and Erie, continued with its outlet 

 flow ing past Chicago to the Des I'laines, Illinois, and .Mississippi 

 rivers, while the country including lake Superior, the northern 

 part of lake Huron, and lake Nipissing, that is, the whole 

 northern side of lake Warren, was uplifted about 350 to 450 

 feet along its extent of 600 miles from east to west. The 

 existence of lake Warren was terminated by the recession of 

 the ice-sheet from the area between lakes Erie and Ontario, 

 when the Niagara river began to flow and to channel the gorge 

 six mile.s long below its receding falls, from which the cmnputa- 

 tion for the time since the Ice Age is derived. The Niagara 

 gorge measures the time after the outflow p.ist Chicago ceased, 

 lake Warren being then succeeded in the basins of the upper 

 lakes, above Erie, by the glacial lake Algonquin, while in the 

 Ontario basin the ice-bound lake Iroquois outflowed past Rome, 

 N. v., by way of the Mohawk and Hudson to the sea. 



Seven-eighths of the difi'erential uplifting which carried the 

 watershed east of lake Nipissing above the level of lake Algon- 

 quin had taken pl.ace before the norlh-castward retreat of the 

 ice-sheet uncovered the Niagara area. Tor some later lime the 

 ice-barrier must have remained upon the Mattawa and Ottawa 

 areas, forbidding any outflow there from lake Algonquin ; and 

 it seems very probable that within that time the continuation of 

 the uplift had raised the watershed .so high that no discharge 

 from the upper lakes ever passed over it. During the ensuing 

 existence of lake Iroquois the Ontario basin was umlergoing a 

 rapid northward uplift, which doubtless was shared by the 

 Nipissing area, so that if any outflow occurred there it must 

 have been very brief, being ended when the land east of lake 

 Nipissing rose higher than the present course of outflow by the 

 St. Clair and Detroit rivers to the Erie basin and Niagara river. 

 The duration of the outlet to the Mattawa could probably have 

 been only a few hundred years, at the longest, if it ever existed. 

 With this po-siblc exception, the present volume of the Niagara 

 river has l>cen maintained during all the time of its gorge ero- 



NO. 1287, VOL. 50J 



