MaV 22, 1890] 



NATURE 



79 



menacing as to resemble the old Caputo. AH of them were 

 roaring, and emitting white fumes. 



The fumaroles of the outer rim, including Caputo, were very 

 active. These latter worked continuously, whilst the new one 

 on the inner edge would stop and start afresh — a phenomenon I 

 have occasionally seen at Vesuvius, in fumaroles which are in 

 direct communication with the lava. The intermittence, then, 

 seems to be due to the surging up of the lava so as to block from 

 time to time the lower inlet, or to be in other cases dependent 

 upon the bursting of the great vapour bubbles as they rise in 

 the viscous paste. 



If this is really the termination of the eruption, we have gained 

 some considerable advance in the interpretation of the eruptive 

 phenomena of a highly acid magma, which is of such feeble 

 character as to be incapable on the one hand of producing a 

 typical pumice, and on the other of giving rise to an outflow of 

 lava. As before stated, differences of opinion will probably be 

 raised as to the nature of the essential ejectamenta, and I have 

 little doubt that it will be dubbed as being more basic than it 

 really is in consequence of the presence of impurities of olivine, 

 augite, &c. It may be wise, therefore, that the reasons that lead 

 me to conclude its acid nature should be given. First and fore- 

 most, we have the intense viscosity indicated by the long inter- 

 vals of the explosions and the bnad-crust structure in the ejecta- 

 menta. Secondly, these bread- crust bombs I have only met 

 with in the ejectamenta accompanying either rhyolitic or 

 trachytic glassy eruption, such as the obsidians of Rocche Rosse, 

 Forgia Vecchia in Lipari, and Monte Rotaro in Ischia. In 

 the former locality we have a beautiful illustration of the forma- 

 tion of these bombs outside the crater. Towards the end of 

 the Rocche Rosse explosive stage, during which the great crater 

 was drilled and the while pumice erupted, a large mass of ob- 

 sidian was hurled up, and fell on the crater edge at Monte 

 Pelato. In consequence of the sudden shock on reaching the 

 ground, the semi-plastic mass cracked, and each fragment, 

 relieved from the surrounding pressure, expanded into a small 

 bread-crust bomb. 



In the third place, the glass of these Vulcano bombs is ex- 

 ceedingly light and transparent, and indicates anything rather 

 than an abundance of any basic iron silicate. 



On looking back through the records of fairly well described 

 eruptions, I cannot resist the impression that the duration of an 

 eruption, other proportions being maintained, is in direct ratio 

 to the basicity of the magma which in fact brings about such a 

 result in consequence of the higher viscosity as the proportion 

 of SiOj increases. Of course more or less advanced crystalliza- 

 tion will also have an influence, as well as the relative higher or 

 lower temperature, in eruptions of pure glass, beside the greater 

 or less abundance of dissolved water. 



The appearance of so many new fumaroles which we did not 

 see six months ago all indicates that Vulcano tend-; (provided 

 j there are no more active signs) to pass into a solfalaric stage 



' such as is its usual state. 



In fine, I must thank Mr. L. Sambon, for so kindly observing 

 carefully the phenomena at Vulcano and transmitting to me his 

 notes, and also Mr. J. P. Iddings for information on the same 

 subject. H. J. Johnston-Lavis. 



Naples, April 18. 



Panmixia. 



I AM glad to observe that his private correspondence has led 

 Prof Lankester to regard the doctrine of "panmixia," or 

 "cessation of selection," in a much more favourable light than 

 heretofore. 



The form in which I stated this doctrine in 1874, and again 

 in the present correspondence, is the form in which it has like- 

 wise been stated by Mr. Galton in 1875, ^Y ^''o'^- Weismann in 

 several of his essays during the past decade, and by Mr. Poulton 

 in his recent lectures. But, speaking for myself, I can see no 

 objection to the form in which it is now presented by Prof. 

 Lankester. For it seems to me immaterial whether we say -that 

 panmixia leads to a degeneration of size, shape, or structure, 

 because the previously sustaining power of selection has been 

 withdrawn ; or whether we proceed to say that the reason why 

 selection has a sustaining power is because, so long as it con- 

 tinues operative, its operation consists in eliminating variations 

 below the standard of full efficiency. But although it appears to 

 me that the latter point goes without saying, if its expression 

 changes the whole aspect of the case in the view of Prof. 



NO. 1073, ^'OL. 42] 



Lankester, I can only regret that I did not express it in the first 

 instance. I did not, however, understand that there was any 

 question touching the fact of variations occurring below the 

 standard of full efficiency, even as regards fully-developed organs 

 of " well-established species." Therefore my argument was 

 directed to show that, upon the " assumption " of such variability, 

 under cessation of selection the standard will not rise above the 

 previous "selection-mean," but always tend to fall below it, 

 on account of reversion, &c. 



Obviously, however, if we disallow that selection has any 

 sustaining power, the doctrine of degeneration as due to its 

 cessation becomes "absurd." Or, which is the same thing, if 

 we "eliminate altogether" the "assumption" of congenital 

 variations occurring below the standard of full efficiency (when 

 once the parts in question have been completely developed by 

 natural selection),and if we substitute a logically '^possible " denial 

 of such variations in respect of such parts by " assuming the 

 ratio of birth-mean and selection-mean to be one of equality " — 

 then, indeed, "the point of interest shifts." But surely the 

 burden of proof lies on the side of anyone who denies this 

 variability to fully-evolved organs. Even in the case of " well- 

 established species " it is " improbable that there is identity 

 between these two means " — or, in other words, that when once 

 an organ has been fully evolved by natural selection, it no longer 

 requires to be maintained by natural selection. 



Again, "that some cases must occur in which the selection - 

 mean -size is [actually] smaller than the birth-mean-size," 

 appears to me true only of cases in which selection has been 

 reversed — as, for instance, in flightless insects of oceanic islands. 

 In such cases natural selection is actively engaged in pulling 

 down its previous work. If natural selection be then withdrawn 

 altogether, the adult-mean-size will probably increase. For not 

 only will there now be no reversal of selection, but cessation of 

 the iieiver selection will enable atavism in some measure to re- 

 establish the state of matters which previously existed under the 

 older selection. Such, at any rate, are the only cases in which 

 I can imagine even the abstract "possibility" of the cessation 

 of selection leading to an increase inr^ize. 



In short, the cessation of selection must always produce the op- 

 posite results to those which were produced by the selection which 

 has ceased — unless, of course, there be any cases in which there 

 is an " identity between the birth-mean and the selection-mean" 

 [i.e. an absence of specific mutability). But even as regards such 

 cases, if they are "assumed" to occur, the assumption amounts 

 to a begging of the question by supposing that the selection has 

 already ceased, and ceased when the parts had reached the point 

 of their maximum development — an assumption which requires 

 to deny any further mutability in respect of such parts, and 

 therefore seems to me well-nigh incredible. Nevertheless, I 

 fully allow that the more " well-established " — i.e. the less vari- 

 able — a species, the smaller will be the necessity for the main- 

 taining power of selection, and hence the smaller effect will 

 result from its withdrawal. This, indeed, we see to be the case 

 even in our domesticated animals — the 'inflexible" goose, for 

 instance, having suffered less change at the hands of panmixia 

 than any of our other farm -yard animals. ^ 



' Nearly all our other domesticated animals yield abundant proof of the 

 potency of panmixia (witness the care with which _" methodical selection " is 

 practised on the progeny of pedigree strains), and if we distrust the analogy 

 between artificial and natural selection in this case, we seem to be rather 

 aiming a bluw at the principflevi lence of the whole Darwinian theory. But 

 as panmixia must act more rapidly, and more completely, in the case of such 

 newly-acquired products of heredity than it is likely to act in wild species. I 

 agree that experiments ought to be tried upon the latter. Moreover, I fully 

 accept the distinction which Prof. Lankester has drawn in his letter of the 

 ist inst. between "size" and "structure." But I may remark that the 

 effect of this distinction is not to indicate that panmixia will have no power 

 to reduce size, while it is capable of entirely abolishing structure. What it 

 does indicate is, that because there are greater potentialities of variation in 

 the case of " complex " structures than in the case of mere "bulk," the sus- 

 taining power of natural selection is of correspondingly more importance : 

 hence the cessation of selection will lead to the disintegration of structure 

 more rapidly and more completely than it will to the reduction of bulk — as I 

 have already pointed out elsewhere in relation to the eyeless peduncles of 

 dark-cave Crustacea. Touching other minor p lir.ts, I may further remark 

 that while in his earlier letters Prof. Lankester accepted Darwin's view that 

 parts are highly variable when selection is withdrawn, in his letter of May i 

 he says it is ''incontrovertible" that the "only effect" of su:h wit drawal 

 must be to increase the number of individual-! near the average mean— /.^. 

 that panmixia \kiI\\ permits and prevents variability. Again, witS regard to 

 what he says about there being no proof that the economy of growth is 

 absent in highly-fed domesticated animals, see Darwin, " Variatiop,"_&c, 

 vol. ii. pp. 345 46, where the best imaginable " proof" of this fact is given. 

 And with regard to his criticism on my use of the terms "essence" and 

 " cause." see Mill, " Logic""vol. i. pp. 53 and 378 et se,/., where the popular 

 abuse of both these terms is shown to be exactly that which 1 have avoided. 



