Alternating generations; a biological study of oak galls and gall flies

Front matter

/ilterpatipi^ (Jepcratiops

a Study of

0^\ Qalls apd

Chapter I. Early Observations and Methods of Research i

II. Descriptions OF Cynipidae Observed. Table

of Alternating Generations ... 9

III. On Gall Formation ..... 97

IV. The Ovipositor and the Egg . . .110 V. Grouping of the Cynipidae . . . 129

VI. Alternating Generation, and Cyclical

Propagation ...... 149

Pediaspis aceris and Bathyaspis aceris . . . 159

Cynips Kollari ...... By the Editor 163

Synoptical Table of Galls .... „ 168

Classification of the Cynipidae ... „ 172

Introduction

A GALL is an abnormal growth of plant tissue produced by animal agency acting from within. All the natural orders of plants include species which are liable to be made use of by insects in this way. Each is visited by its own special gall-maker, which need not necessarily belong to the Cynipidae, for gall-makers are also found among the Coleoptera, Lepidoptera, Diptera, Nematoda\ and in other classes. Any organ of the plant may become the seat of this hyperplasia, but the form which the gall ultimately assumes is governed by the potentialities of growth in the part attacked, and by the nature of the animal excitation present. The rose and some Compositae produce well known galls, but the oak is the favourite home of the Cynipidae. In this monograph Dr. Adler has described those oak-galls and gall-flies most commonly found in this country, with the exception of Cynips Kollari, the Devonshire marble-gall, which does not occur in Germany north of the Elbe ; as it is however one of the most familiar galls on English oaks, a description of it has been added in the appendix. Before Dr. Adler

' H. Charlton Bastian, ' Monograph of the Angulllulidae,' Lin. Soc. Trans, vol. xxv, 1866.

X Introduction.

had demonstrated the existence of cycHcal propagation, many curious explanations were offered, in order to account for the lengthened interval that elapses between the death of one generation and the appearance of the next. The currant-gall, for example, appears on the male catkins of the oak in May, the fly quits the gall in June, lives for a few days only, and nothing more is seen until the male catkin appears again next May. What had become of the eggs in this long interval ? Spontaneous generation of the insect, within a gall that had no external opening, had its advocates. Later it was believed that a form of metempsychosis took place, and galls were among the stepping-stones in the path of organic evolution, by which the vegetable passed into the animal soul. By some it was supposed that the eggs, found in the fly in June, reached the ground, whence they were drawn up, mingled with the sap, and arrested next spring in the leaves or flowering catkins, there to form the currant-galls again. Dr. Adler, by proving the existence of cyclical propagation, has shown that the interval between the appearance of the currant gall-fly in one spring, and its reappearance in the next, is occupied by another agamous generation. But while he showed that this rule holds good for the majority of species, he has also demonstrated that, in some at least, no sexual generation now exists.

Pliny ' knew that a fly was often produced in a gall, but he did not associate it with the cause of gall-growth ; on the contrary, he thought galls grew in a night, like fungi. Many early observers, however, considered them as insect productions and were aware that a variety of insects emerged from them ; but the attention which some of these authors bestowed upon this subject was

^ Pliny, Nat. Hist. xvi. 9, lo ; xxiv. 5.

Marcellus Malpighi. xi

not always due to its biological interest. Mathiolus, one of the best of the commentators of Dioscorides, and a believer in spontaneous generation, declared that weighty prognostications as to the events of the year could generally be deduced by observing whether galls contained spiders, worms, or flies. Most of the older writers describe gall-flies, which are now known to be agamous, as possessing males ; but their descriptions are often perfectly clear, and the flies can be recognized as the males of one of the Torymidae, or of some other species of parasite. There is no reason to think that any males of agamous species were actually in existence at the time when these authors wrote.

The earliest systematic writer on galls was Marcellus Malpighi, Physician to Innocent XII. He was Professor of Medicine at Bologna, and afterwards at Messina, and his treatise ^ ' De gallis,' published in 1686, gives an extremely accurate description of the galls then found in Italy and Sicily. Dr. Derham, Canon of Windsor, in the notes to his Boyle Lectures •(1711-1712) compares Malpighi's account with the galls then found in England, and says, ' I find Italy and Sicily more luxuriant in such productions than England, at least than the parts about Upminster (where I live) are. For many, if not most, of the galls about us are taken notice of by him, and several others besides that I have never met with, although I have for many years as critically observed all the excrescences and other morbid tumours of vegetables as is almost possible, and do believe that few of them have escaped me ^.'

Malpighi's work does not appear to have been known either to Linnaeus or Fabricius ; they include

' Dioscorides, i. 146, Paris, 1549.

* W. Derham, F.R.S,, Physico-theology , iii. c. 6.

xii Introduction.

inquilines and parasites under the genus Cynips, indeed St. Hilaire, Latreille, and Olivier reserve the name Cynips for certain Chalcididae, and use Diplolepis for the true gall-maker, but with few exceptions subsequent writers have applied the general name Cynips to the gall-makers. It is well, however, to bear this change of nomenclature in mind, when the males of certain species are said to have been found \ Reaumur has left excellent descriptions and drawings of many species of galls '", but the first to bring order out of the confusion in which the Cynipidac still remained, was Theodor von Harlig of Brunswick^. He greatly improved the existing classification of this family and carefully pointed out the true relationship which the various dwellers in galls bore to each other. He arranged gall inhabitants into three classes ; first, the gall-makers which he compared to the actual householders ; secondly, inquilines, guest-flies, cuckoo-flies or lodgers, who take up their quarters uninvited within the gall, and live on its food stores, but do not directly aim at the gall-maker's life ; and thirdly, parasites who deposit their eggs on the larvae of their host or his lodgers, with the object of destroying them, and who are therefore murderers. Besides those living in a state of symbiosis, there are also true commensals in some galls.

It is the simultaneous presence of these various classes that frequently gives rise to confusion in carrying out breeding experiments. Synergi among inquilines resemble true gall-making species so closely, that caution

' Westwood, Zoological Journal, No. 13 ; Cameron, P., British Phyt. Hymen, vol. iii. p. 140.

'' Reaumur, Me'nioires pour servir d Vhistoire des insectes, 1734-42.

■ Hartig, ' Ueber die Familien der Gallwespen,' Gerniar^s Zeitschr. f. d. Ent. II. Heft i. p. 176, 1840; III. 322-358, 1841 ; IV. 395, 1843.

Gall- Dwellers. xiii

is always necessary to see that the gall-maker, and not a Synergus, has been obtained \ Both Malpighi and Canon Derham were aware of the attacks of parasites, and actually saw galls pierced by them. The latter says, ' I apprehend we see many vermicules, towards the outside of many oak-apples, which I guess were not what the primitive insects laid up in the gem from which the oak-apple had its rise, but from some supervenient additional insects, laid in after the apple was grown, and whilst it was tender and soft'.' Ratzeburg, a forester like Hartig, in his beautiful * Forstinsekten ' " corroborated Hartig's division of gall-dwellers. Giraud, Schenck, Reinhard, Taschenberg, Schlechtendal, Wachtl, Forster, and Lichtenstein, have since each advanced our knowledge of the Cynipidae, and the history of galls generally has been admirably written by Lacaze-Duthiers. The entomologists of America have not been behind those of Europe ; Baron C. R. Osten- Sacken, before he quitted America in 1877, had discovered eighteen new species ; Bassett, thirty species, and Walsh and Riley had each added much to our knowledge. Professor Gustav Mayr of Vienna has not only increased largely the work of previous observers, but has arranged all that is known of gall-makers and gall-dwellers in a series of admirable monographs * and has

' See Walker, But. Mag. vii. p. 54.

' Derham, Physico-theology, iii. p. 389.

' Ratzeburg, Die Forstinsekten, vol. iii, Berlin, 1844.

* Mayr, G., Die initteleuropdischen Eichengallen in Wort und Bilderu, Wien, 1870-71 ; Die Einmiethler der niitteleuropdischen Eichengallen, Wien, 1872; Die europdischen Cynipiden-Gallen tnit Ausschluss der auf Eichen vorkoynmenden Arten, Wien, 1876 ; Die europdischen Toryniiden . Wien, 1874; Encyrtiden, 1876; Olinx, 1877; Eurytonta, 1878: Telenotniis, 1879; Die Genera der gallenhewohnenden Cynipiden,'W\^n, 1881 ; Die europdischen Arten der gallenhewohnenden Cynipiden, Wien, 1882.

xiy Introduction.

drawn up excellent analytical tables of species. In this country Westwood, Halliday, Walker, and Fitch have all done good work, and Mr. Peter Cameron^, in his account of the Phytophagous Hymenoptera, just completed for the Ray Society, has brought together an immense amount of valuable material, new and old.

Dr. Adler began his observations on gall-flies in 1875, and added certain new species described in these pages. His important work, however, was the unfolding of their life-history ; proving that while many species were linked together in alternate agamous and sexual generations, others remained wholly agamous.

Among biologists theories of dimorphism and metagenesis had been discussed in connexion with the Cynipidae and were current as early as 1865 '^ ; in 1873 Bassett announced his theory of seasonal dimorphism^ ; and in the same year Riley established the existence of alternation of generations between Cynips operator and Cynips opcratola\ The numerous publications of Thomas and Frank have cleared up many special points, and Beyerinck ^ in an admirably illustrated monograph, full of original work and careful reasoning, has added much both from a botanical and zoological point of view.

The existence of alternating generations in living

' Peter Cameron, A Monograph of the British Phytophagous Hymenoptera, 4 vols. 1882- 1893. Ray Society.

'' Reinhard, ' Die Hypothesen iiber die Fortpflanzungsweise bie den eingeschlechtigen Gallwespen,' Berlin. Ent. Zeitschr. vol. ix. 1865.

■' Bassett, Canadian Entomol. vol. v, pp. 91-94, May, 1873.

* Riley, American Naturalist, vol. vii. p. 519, 1873.

' Beyerinck, Dr. M. W., ' Beobachtungen iiber die ersten Entwicklungsphasen einiger Cynipidengallen,' Natnurk. Verh. der Koninkl. Akademie, Deel xxii.

Cyclical Propagation. xv

organisms was first discovered by Chamisso', the author of Peter Schlemihl,' who in 1815 accompanied the Chancellor Rumjanzow's expedition as naturalist in a voyage round the world. He noticed that among the Salpae a solitary salpa gave rise to a generation of a different form, united in chains of twenty or more, and that each link of this ' associated ' form again produced the ' solitary ' form. He concluded that all solitary salpae produced chains, and on the other hand that all those in associated chains were the parents of solitary ones ; so that a salpa mother was not like its daughter or its own mother, but resembled its granddaughter and its grandmother. At first the accuracy of Chamisso's observations was doubted, chiefly for the reason that no similar phenomenon was then known in nature. By degrees, however, facts began to accumulate ; in hydroids and flukes similar generation-cycles were observed, and in 1842 Steenstrup^ collected all that was known on the subject of alternating generations into a monograph published in Danish and subsequently translated into German and English. He described this mode of reproduction as ' a peculiar form of fostering the young in the lower classes of animals.'

In 1849 the late Professor Owen^ suggested the existence of a residual germ-force in the cells of the apparently asexual generation, and thought that alternating generations were due * to the retention of certain of the progeny of the primary impregnated germ-cell, or in other words to the germ-mass, with so much of the spermatic force inherited by the retained germ-cells from

' Adalbert de Chamisso, De Animalibus, Fasc. i, De Salpa, Berlin, 1819.

^ J. J. S. Steenstrup, Ueber den Generationswechsel, Kopenhagen, 1842, and Ray Society, T845.

^ R. Owen, Parthenogenesis, 1849.

xvi Introduction.

the parent cell or germ-vesicle, as suffices to set on foot and maintain the same series of formative actions as those which constituted the individual containing them ' : and this may be taken as the earliest suggestion of the continuity of the germ-plasm.

Fresh instances of this class of phenomena have steadily accumulated, in which the life-cycle of the species may be represented by two or more generations, differing in form and organization, existing under different conditions, and reproducing themselves in different wa3's.

The simplest cyclical rhythm occurs in reproduction by metagenesis, where a sexual and asexual form alternate ; this is the law of development in Medusae and Trematoda. In heterogenesis a sexual and an agamous generation alternate, and in this rhythm the agamous may be juvenile or adult ; in Cecidomyia, for example, the parthenogenetic generation reproduces itself when still immature, while in the Cynipidae, on the other hand, it does so only when perfectly developed. Another variety occurring between one hermaphrodite and one sexual form, is seen in Angiostomiirn nigrovenosum, the lung parasite of the frog ; and a still more perfect alternation is found in the thread worm of the snail, Leptoptera appcndiculata, where two perfectly formed sexual generations are linked in a cycle.

Sometimes the sexual or asexual member of the cycle may be complex. The liver-fluke of the sheep gives rise to an active ciliated aquatic embryo, which, after a time, pierces and enters a water-snail to become a passive sporocyst; from its germ-cells fcdiaesLve formed within the sporocyst, and after several asexual generations, they give rise to minute ccrcariae, which leave the snail and creep up the stalks of grass ; here they

The Larval Theory. xvii

become encysted, are eaten, and grow within the sheep to become adult sexual flukes. In this series the Cercaria and fluke form members of the sexual division; all the others are members of the asexual division of the cycle.

In all C3'clical propagation, whether in the animal or vegetable kingdom, there is a tendency for one generation to become subordinated to the other. In flowering plants the sexual is subordinated to the asexual, and even some ferns exhibit a similar tendency, a fernplant rising vegetatively from the prothallus ; while in other ferns there is a tendency to apospory, a fernprothallus springs from the site of the spores, and the asexual becomes subordinated to the sexual. In flukes, in the same way, rcdiac ma}^ be budded off frofn the sporocyst and the species be continued without ever attually becoming sexual. In the Cynipidae it will be seen that in some, like Cynips Kollari, the sexual generation has been wholly subordinated to the asexual, and in others, like Rhodites rosae, this process is still going on and males are becoming functionless and extinct.

Leuckart^ regarded interpolated generations as larval states, and following his teaching, Lichtenstein ^ looked upon the agamous as larval stages of the sexual species. He believed that the biological evolution of a gall-fly extended from the time when a female emerged from a true egg in a condition to be fecundated by the male, until another egg was reached presenting the same conditions. All other stages he considered as larval, although in them reproduction by budding was possible, and he held that in this way insects might go on reproducing themselves indefinitely without ever reaching

- J. Lichtenstein, Les Cyutpides, 188 1, p. x.

b xviii Introduction.

the perfect state. Certain Phylloxeridae reproduce themselves by subterranean budding, without ever arriving at the winged and sexual forms : he regarded this as analogous to the multiplication of plants by suckers, without the intervention of seed. He compared the life-history of gall-flies with that of Phylloxera, and gave it as his opinion that, among the gall-flies, the individuals o^ Neuroterus lenticular is had been mistaken for females, because they possessed an ovipositor and eggs, that, properly speaking, they had no sex, but were only larval forms of Spathegaster baccarum, their so-called eggs being gemmations.

It is difficult to regard species which have a perfect female form, and ovaries filled with perfect eggs, as being perpetually in the larval state, because they are wholly agamous and have no males. In Rhodites rosae, as has been said, the few males found are functionless, and are disappearing; it seems therefore more in accordance with facts to speak of the individuals of agamous species as females, the males of which have ceased to appear. Although the Aphides are scarcely comparable to the Cynipidae, since they possess a form which is asexual, they may nevertheless be arranged, like the Cynipidae, in alternate generations, without having recourse to the larval theory.

The importance of sexual reproduction has been greatly exaggerated, and there seems no good ground for assuming that a generation ought to extend from one fecundated egg to another. In Cynips Kollari, and many others, no sexual generation is known ; the flies all possess perfect female forms, but they are agamous and have the power of reproducing themselves by parthenogenetic eggs. At the same time it is difficult to believe that the agamous can be the primitive

Advantages of Parthenogenesis. xix

form ; or that perfectly formed sexual organs could have been evolved unless the sexual had been the earlier generation.

The power of producing parthenogenetic eggs is widely distributed among the arthropods, and appears to come into operation whenever it secures the existence of the species more effectually than sexual reproduction. When at one season of the year sexual, and at another agamous, reproduction is the more beneficial, then heterogeny will be found to prevail. In one group amphimixis has been wholly abandoned, but its members are enormously prolific, and their eggs have the power of resting over more than one year. By means of partial parthenogenesis, a much more rapid increase is ensured than could have been possible, in the same time, by sexual reproduction only. Every individual of the winter generation, unhindered by the requirements of fertilization, is engaged in laying eggs ; the number of the sexual individuals hatched from these eggs is consequently enormously greater than it could have been, had only half the winter generation been of the female sex, and had that half, in order to be fruitful, been dependent on the chances of fecundation. When generations alternate, there are alternating advantages to the species. The winter generations emerge from the gall at a time when pairing is not easy, and it is a distinct gain to the race when every individual has the power of reproducing itself independently. The summer generations, on the contrary, appear in halcyon days, when there is nothing to mar their nuptial flight, and then the species obtains greater variation, and those physiological advantages of amphimixis which parthenogenesis cannot afford.

But amphimixis is in no way essential to heredity. b2

XX Introduction.

The oak gall-flies about to be described will be found to afford examples of heredity occurring among purely agamous species, as well as among those which alternate between an agamous and a sexual generation. The characters of each generation are inherited and passed on in two perfectly distinct streams. The advance which has in recent times been made towards a clearer comprehension of heredity, is in great measure due to the influence of Weismann ', who, by discarding the idea that sexual reproduction is in any way fundamental or essential to life, has led us to regard the facts of heredity wholly untrammelled by it. Amphimixis undoubtedly has its advantages, but descent may be continuous in the female line, or there may even be a male parthenogenesis ^

To understand the mechanism by which alternation of generations is brought about, it is necessary to recall the minute structure of the sexual cells, and especially the behaviour of their nuclear contents ; and to observe the difference in the extrusion of the polar cells, as occurring in the eggs of parthenogenetic and sexual species.

The changes in the ovum during nuclear division are briefly these '. After a resting stage the chromatin granules, which there is reason to believe are the material bearers of hereditary qualities, appear as a thread, apparently, spirally coiled within the nucleus. An accessory nucleus forms and divides into two ; at each pole of its achromatic spindle is placed a centrosome with its surrounding attraction-sphere ; the nuclear membrane disappears ; the chromatin

' Weismann, Essays on Heredity, vol. ii. p. 86, 1893.

- Ibid. vol. i. p. 253 ; Schenck, Handbitch dcr Botaitik, Bd. ii. p. 219.

^ Weismann, Essays on Heredity, vol. ii. p. 118.

spermatogenesis. xxi

granules become aggregated into rod-like chromosomes of equal length and of constant number in each species ; these are formed at the equator of the spindle, and split by longitudinal fission, so that their number becomes doubled. One star of chromosomes is drawn to the centrosome at each pole of the spindle, and thus two daughter-nuclei, which for convenience I will term oocytes, are formed, of unequal size ; the smaller one being extruded as the first polar body. In the parthenogenetic egg this completes the division, but in the sexual egg a second nuclear division follows immediately on the first without a resting stage ; the oocyte divides into two oozoa, and one oozoon, containing half the chromosomes, is extruded as the second polar body. The oocyte which forms the first polar body is observed to split into two oozoa after extrusion.

In this way three of the oozoa which have arisen from the division of the primitive germ-cell have become polar bodies, while the remaining oozoon, containing one half the number of chromosomes which the primitive germ-cell contained, is left in the nucleus, functional or capable of development.

Within the spermatic tubes of the male, corresponding changes take place in the primitive sperm-cell. Its chromosomes are doubled and it divides into two spermatocytes (spermatogens), each of which again divides without a resting stage into two spermatides (spermatoblasts). The four spermatides undergo various changes during which they become elongated, the nucleus, containing the chromatin elements, becomes the head of a spermatozoon and ends in a motile barb. The protoplasm of the cell-body is drawn out into a sheath through which the filament passes, and is continued beyond as the vibratile tail.

xxii Introduction.

We have therefore this arrangement of parts in the male and female cells, to which for convenience of comparison I will apply these names.

I. Primitive germ-cell. = Primitive sperm-cell. II. Two oocytes of = Two spermatocytes, which one forms the first polar body.

III. The oocytes dividing = The spermatocytes diinto four oozoa, of viding into four sper-

which two are pri- matides, each becoming

marily functional a spermatozoon, two of

in the partheno- which may become re-

genetic and one in productive in a male

the sexual egg. parthenogenesis, one is

primarily functional in the sexual egg.

There seems little doubt that the chromosomes containing the germ-plasm of the species, are identical in the maternal and paternal reproductive cells, and that so long as their number is complete or sufficient to enable conjugation to take place, it is immaterial from which parent they may come. Boveri ' succeeded in denucleating the egg o^ Echinus microtuberculatus and introducing spermatozoa of another species, Sphaerechinus granularis ; when the egg developed, the larva was found to belong to the latter species, although living on the vitellus of the former. This was a case of male parthenogenesis. It however appears certain that one oozoon or spermatozoon requires conjugation with one other oozoon or spermatozoon before development can take place.

* Boveri, ' Ein geschlechtlich erzeugter Organismus ohne miitterliche Eigenschaften, ' Gesells./. Morph. u. Phys. Munchen,Ju]y i6, 1883.

The Germ-Plastn. xxiii

According to Weismann, the chromosomes are 'idants,' formed of smaller groups called 'ids'; these are made up of ' determinants,' which are again composed of ' biophors ' or ultimate units. The biophors are more or less equivalent to the ' physiological units ' of Herbert Spencer, the 'plastidules* of Haeckel, the ' gemmules ' of Darwin, and the ' pangenes ' of De Vries ; although these variants are not exactly analogous to each other, they are all ultimate molecules of inherited plasm ^. The nuclear thread is, so to speak, the web of the individual's destiny, and the ids which it contains are the inherited working plans of its architecture. As development advances, the ids are disintegrated into determinants, and the determinants into biophors, each group getting smaller until every biophor ultimately reaches and controls its own cell. The way in which the germ-plasm assumes the direction of ontogeny resembles a battalion marching to outpost duty. The main body steadily proceeds to distribute itself according to a predetermined plan ; the battalion splits into half-battalions, the half-battalions into companies, the companies into pickets, and the pickets tell off their sentries, until each man at last finds himself at the post he has to guard. And just as a line of communication is always maintained through the battalion, so each nucleus is kept in touch with its cell, and each cell with the cells around it, by means of a fine protoplasmic reticulum which passes through the cell-walls'-. It is obvious that influences affecting any stage of development must affect larger or smaller groups of determinants, just as any command given by an officer will

^ Weismann, Germ-plasm, p. 41.

^ Sedgwick, Monograph of the development of Peripatiis capensis, p. 41.

xxiv Introduction.

affect the number of men particularly under him, and may alter their original course.

In cases of alternation of generations, and in those cases in which the two generations do not differ from each other in the full-grown condition, but in which the eggs develop differently, Weismann holds that there must be two kinds of germ-plasm, each containing the determinants of one form, and that these two must be passed separately along the germ-tracks, from one generation to another, so that each must always contain the other, stored away in an inactive condition \

In the case of gall-flies there is often a well-marked difference of structure between the females of the two alternating generations. The wing determinants in the winged Trigonaspis crustalis (Plate II, fig. i8 a) must have been modified in the wingless Biorhiza renum, (Plate II, fig. i8), but they appear again in Trigonaspis crustalis. It is improbable that lost determinants would be redeveloped in this way, and therefore there must be two alternating sets of determinants. These, Weismann holds, are contained within the same germ-plasm, but never become active at the same time ^

I believe that in the function of the polar bodies we possess the simplest explanation of the problem of alternating determinants in the Cynipidae.

According to Weismann's earlier views ^, the first polar body removed ovogenetic nuclear substance which, after the maturation of the egg, had become superfluous and injurious ; while the second polar body removed as many different kinds of idioplasm as were afterwards introduced by the sperm-nucleus.

He has abandoned this view of the function of the

' Weismann, Germ-plasm, p. 176. ^ Ibid. p. 178.

^ Essays on Heredity, vol. i. p. 365, Oxford, 1891.

Function of Polar Bodies. xxv

first polar body, and, according to his later views, the extrusion of both polar bodies is a provision for withdrawing promiscuously certain ids of germ-plasm for the purpose of securing variation. He says : 'I consider this remarkable and apparently useless process of the doubling and two subsequent halvings of the idants, as a method of still further increasing the number of possible combinations of idants in the germ-cell of one and the same individual ' ' ; the three quarters of the mass of germ-plasm which passes into the polar bodies, he regards as being 'again lost",' while what is left in the ovum, after being increased by one spermatozoon, is halved by the first nuclear division ; one half going to direct ontogeny, and the other half going to form the primary sexual cells : but as these last divide and behave at first like other somatic cells, they must, he thinks, possess active idioplasm as well as unalterable germ-plasm, and they must therefore contain more ids in their nuclear matter than do somatic cells ^

I incline to the view that these cells in the Cynipidae remain somatic, and simply form the rudiments of the sexual organs, but that the function of the polar bodies is to provide the primitive reproductive cells. These bodies contain the inactive germ-plasm, historically prepared for this very purpose, and it is scarcely possible to imagine that nature would create this phylogenetic epitome of the species to be extruded as functionless, or to be broken up and again lost *. It is undeniable that they are of great value in promoting variation. The first division of the nucleus into two oocytes is not a reducing division, but results, so far as the germ-plasm is concerned, in two equivalent cells.

' Weismann, Gerni-plastn, p. 247. - Ibid. p. 248.

xxvi Introduction.

In the sexual egg, however, the second division follows immediately upon the first, and the extrusion of one oozoon reduces the chromosomes by half their number, which is again made good by those of one spermatozoon. While, therefore, in the parthenogenetic egg those 'material bearers of ancestral qualities' remain the same from generation to generation, they are changed in the sexual generation by each cross fertilization, and in this exists the advantage of amphimixis. For the evolution of species, individual variation is essential, and those individuals who derive half their germ-plasm from one parent, and half from another, each bringing in different lines of ancestors, must present greater variation than those who derive their germ-plasm from one parent and one ancestral line. Sexual reproduction is thus of advantage to the species in so far as it helps to increase inherent tendency to variation, and enables evolution to take place by the steady accumulation of slight beneficial differences '. The random withdrawal of the ids of ancestors by a species of lot, as laid down by Weismann, must be considered simply as a provisional hypothesis ; and it seems, according to Prof Hartog, that mathematically its effect, if it took place at all, would be the opposite of what Weismann expected "^