the cytology of sitophilus [calandra] oryzae (l.), s ... · the cytology of sitophilus [calandra]...

50 Cytologia 17

The Cytology of Sitophilus [Calandra] oryzae (L.), S. granarius (L.), and Some Other Rhynchophora (Coleoptera)1

Stanley G. Smith2

Received March 3, 1952

Abstract

Contrary to the findings of other workers, investigation of two calendrid weevils failed to show anomalies in meiotic behaviour: one had the formula 11 AA+Xyp, the other, 10 AA+neo-XY. In this way alone they differ from other Curculionidae, bisexual species of which, with one exception, may be formulated as 10 AA+Xyp or 10 AA+XO. Additional species examined were: Curculionidae: Sitona lepidus, Sitona cylindricollie, Hylobius congener (10 AA+Xyp), and Anthonomus scutellatus (13 AA+Xyp); Scolytidae: Hylurgops pinifex, Dendroctonus engelmanni, and fps pini (14 AA+Xyp); and Platystomidae: Euparius marmoreus (10 AA+Xyp). It is suggested that the

phylogenetic sequence in the Rhynchophora might be (i) Platystomidae, (ii) Curculionidae Otiorhynchinae, (iii) Curculionidae-Curculioninae in general, (iv) Curculionidae-Calend

rinae, (v) Anthonomus, and (vi) Scolytidae.

Introduction

The two most basic generalizations formulated by Darlington (1932) in his original development of the "New Cytology" are that homologous regions of chromosomes pair in twos and in twos only and that the

maintenance of this paired condition is, except in the autosomes at least of the male of certain higher Diptera, including Drosophila, dependent on the formation and persistence in them until metaphase of chiasmata between the homologous chromosomes. These have been repeatedly assailed (see compilation in Hughes-Schrader, 1943, and review by Schrader,,1948): the first on the basis of reputedly direct evidence; the second as being but one of several functionally different mechanisms (as Darlington himself is admittedly aware, as evidenced by his recognition of such forces as somatic pairing, residual affinity, and terminal attraction).

Noteworthy among the aberrant cases that are manifestly incom

patible with both of the above fundamental propositions is that described by Agar (1911) in the spermatogenesis of the South American lung-fish, Lepidosiren paradoxa. Therein, a dissociation of "parasynaptically"

paired homologues is followed by their reassociation during the course of diakinesis. Hence in this species chiasmata supposedly fail to maintain the chromosomes in association until anaphase, yet, despite their doubleness, which Darlington holds prevents pairing, the homologues

1 Contribution No . 32, Division of Forest Biology, Science Service, Department of

Agriculture, Ottawa, Canada.3 Cytogeneticist

, Forest Insect Laboratory, Sault Ste. Marie, Ontario.

1952 Cytology of Sitophilus [Calandra] oryzae (L.), S. granarius (L), etc. 51

succeed in reassociating prior to metaphase. Later, Hogben (1920) re

ported a similar re-pairing of ex-conjugants in certain species of Hymenoptera. The species investigated were, however, diploid obligatory

parthenogenetic ones, in which oogenesis is terminated at or about the first metaphase; consequently the significance of the secondary coupling, and for that matter the earlier pairing, is obscure.

Comparable and additional anomalies in behaviour have subsequently been reported for the females of the weevil Sitophilus [Calandra] oryzae by Tiegs and Murray (1938). According to these workers a similar temporary separation of homologous chromosomes occurs during diakinesis, followed by the reformation of six bivalents by metaphase. The ensuing reduction is said to be accompanied by a separation of the component chromatids of each chromosome, whereby 12 individual elements appear at late anaphase. By some undetermined means, these reassociate in pairs by the second metaphase (presumably of necessity sister with sister), and again disjoin during the second anaphase to

provide nuclei with the haploid number of six chromosomes. Their student Gunson (1945) claims to have fully confirmed these observations on the female and to have found in the male a complement of five autosomal pairs and a single X chromosome. In the latter sex she found "... a similar dissociation of the homologues is intercalated between diplotene and metaphase I...so that the full number of separate elements, in the male eleven, is reconstituted". She maintains, however, that there is no separation of chromatids during the anaphase of pri

mary spermatocytes.

Earlier, Inkmann (1933) reported that in the allied species S. [Calandra] granaries the female had a diploid number of 12 chromosomes

-the same as in S. oryzae-, but he considered the initial oocyte division to be equational, the reduction in number to six being relegated to the second division. Tiegs and Murray questioned this interpretation of the course of events, because it conflicts with their contention that in S. oryzae an equational first division is only simulated owing to the precocious but temporary separation of chromatids, following reduction, during anaphase.

There are, therefore, a number of points claimed in the cytology of these diverse organisms which, being in conflict with the basic rules laid down by Darlington, cannot be ignored. These points are: (1) the

premetaphase dissociation of homologues and (2) their reassociation by first metaphase (Lepidosiren, S. oryzae, and Hogben's Hymenoptera);

(3) the separation of chromatids during the first anaphase and (4) their subsequent reassociation in time for the second oocyte division (S. oryzae female); and (5) an equational first division followed by a reductional

second division (S. granarius female).4*

52 S. G. SMITH Cytologia 17

A further remarkable feature in the cytological descriptions given for the two species of Sitophilus concerns their chromosome number: this, it will be noted, is 12 for the female of both species and 11 for the male of S. oryzae. Available evidence, derived from a consideration of more than 200 species of beetles, more than half of which were known to me at the time through direct examination and the remainder

through the literature, has been outlined (Smith, 1950), pointing to 9AA+Xyp as the basic formula for the Coleoptera3. It symbolizes the most prevalent type found, consisting of nine pairs of autosomes, a large X, and a relatively minute Y chromosome, these last being associated at first meiotic metaphase at two terminal points of contact in the form

of a parachute (subscript p).Suomalainen (1940 a and b, 1947 and 1949), Seiler (1947), Mikulska

(1949), and Smith (1950) have found the Curculionidae to be an exceptional family in this respect. Disregarding Sitophilus and Anthonomus scutellatus Gyll. to be dealt with later, all the species so far investigated cytologically, about 40 in all, may be symbolized as either 10AA+Xyp or 10AA+XO, or are parthenogenetic polyploids obviously derived from one or the other (see Table I). The counts hitherto reported for Sito

philus therefore occasion surprise, not only in their extreme difference from other known weevils, but also because, so far as is known, the lowest numbers of autosomal pairs in XO-species of other families of beetles are six in two species of Necrophorus, in the Silphidae, and eight in five species of Elateridae (Stevens, 1909; Smith, 1950 and 1952 c).

The several cytological anomalies that have been credited to Sito

philus are of such profound theoretical interest as to render their reinvestigation a matter of first rank importance. Carried out in the light

of knowledge gained from an examination of a large number of other

beetles, such a study might be expected to prove fruitful, not only in

ascertaining whether or not Sitophilus in reality constitutes so flagrant

an exception to the principles laid down by Darlington, but also as a

problem with possibly broad phylogenetic implications.

Materials and methods

Cultures of the grain weevil, S. granarius (L.), and the rice weevil, S. oryzae (L.), were kindly supplied by Mr. H. A. U. Monro, Science Service Laboratory, Department of Agriculture, London, Ontario. Additional material representative of the large and small strains of the latter species (Birch, 1944 and 1946; Richards, 1944) was subsequently

3 Darlington (1937) considers the extreme difference in size of the X and Y chromosomes in Coleoptera as having resulted from progressive diminution of the Y. According to my survey this condition predated the origin of the order; evolution within it has evidently proceeded by increase in the size of the Y chromosome (see Asana, Makino, and Niiyama, 1942; Smith, 1951-contra Guenin, 1950).

1952 Cytology of Sitophilus [Calandra] oryzae (L.), S. granaries (L.), etc. 53

provided by Dr. L.C. Birch4, University of Sidney, Australia, to whom I am greatly indebted. Since no differences were discernible between

the two Australian and the Canadian cultures, these will be described

together in the following section.

Adult males of both species provided an abundance of spermato

gonial and spermatocyte divisions, but, as is usual, the maturation divisions of the females are deferred until after oviposition. As an investigation of the later oocyte stages almost always proves to be a long

term and tedious procedure and often provides only unsatisfactory re

sults, it was decided to restrict observation to males and to the gonia

and pachytene stage of females, unless otherwise warranted. As will

become apparent this decision was fully justified.

An opportunity is taken herein to present the pertinent cytological data concerning four other species of Curculionidae-the family from which Sitophilus has been removed by Boving and Craighead (1931)-three species in the Scolytidae, and in addition a member of the Platystomidae

(determinations by courtesy of Mr. W.J. Brown, Systematic Entomology, Ottawa). These latter two families have not previously been examined cytologically.

Material was prepared by the squash method already described

(Smith, 1943): fixation in modified Kahle's fluid (with, however, absolute

alcohol replacing 95% ethyl alcohol for the species of Sitophilus), and

staining by Feulgen and light-green (16 min. hydrolysis at 60•Ž.; 1 to

2 hr. Feulgen). Additional check slides were made with acetic-orcein.

All drawings were originally made at a magnification of approximately

•~ 6,600; in reproduction they are reduced to about •~3,300. The biva

lents from primary spermatocytes at metaphase have been aligned and

rearranged according to size, except that the sex-determining pair is

invariably placed at the extreme right. This procedure not only facili

tates comparison but allows of economy in space.

Observations

Sitophilus granarius:

Both males and females have 24 chromosomes in gonial metaphase

(Text figs. 1 and 2). That this is the true diploid number and not the result of counting a double set of daughter chromosomes at very early anaphase (v. Gunson, 1945) is shown, first, by the presence of 24 chromosomes in spermatogonial prophases and, second, by the split condition of the chromosomes at prometaphase in both, sexes (see Text fig. 8 for a comparable stage in S. oryzae). There is considerable range in the

4 I am further indebt to Dr. Birch for drawing to my attention that Richards

(1944) had earlier credited Mr. F. C. Grigg with establishing that both the large and small strains of S. oryzae had 2n=22 chromosomes.


size of the chromosomes, which renders it difficult to distinguish both

sex chromosomes from the autosomes in spermatogonia, but the y is

presumably the small, spherical one.

At pachy

tene approxi

mately 12 biva

lents are pre

sent in both the

male and the

female. In the

former one

pair, presumably the sex

determining

pair, is positively hetero

pycnotic. The heteropycnotic

bivalent lies in

close associa

tion with the

nucleolus, a

situation which

is taken to sup

port the inference that it is

the sex-deter

mining pair.

In the female

no differen

tially condens

ed bivalent is

evident. Post

pachytene stages were not

examined in

the female.

Text figs. 1-7. Sitophilus granaries: 1 , spermatogonial metaphase, 24 chromosomes including spherical y; 2, oogonial metaphase

, 24 chromosomes one of which shows separation into chromatids; 3

, male diplotene with about 12 bivalents-it is not possible to dis

tinguish chiasmata from overlaps in all cases; 4, male diakinesis , 12 bivalents; 5-7, first spermatocyte metaphases showing 11 auto

somal bivalent and the Xyp sex-determining pair.

The conven

tional sequence

of later pro

iphase stages is

encountered in males, passing through diplotene, where approximately

the haploid number of bivalents can be counted (Text fig. 3), and

1952 Cytology of Sitophilus [Calandra] oryzae (L.), S. granarius (L.), etc. 55

diakinesis, where there are 12 bivalents (Text fig. 4), to the final con

gression of the bivalents on the equatorial plate. At first metaphase, the components of the 12 bivalents are in all cases associated terminally

giving either dumbbell-like or ring-shaped configurations. They differ

greatly in size (Text figs. 5 to 7), as is to be expected from the wide range in size so evident in gonial metaphases. Of the three bivalents that are particularly small, one, probably the smallest, usually appears to be slightly negatively heteropycnotic as well as heteromorphic. By

these criteria, this is the Xy pair, but the X member is so nearly the same size as the y chromosome that the great disparity in size of the sex chromosomes of typical Xyp-species of beetles (see Text figs. 17-25, excluding 19 and 21) is not available as a diagnostic character. Also lacking, presumably on account of the more equal size of its components, is the delay in congression characteristic of the Xyp bivalents of species such as Sitona lepidus (Text fig. 17). The first division is numerically reductional, 12 chromosomes passing to each pole. At late anaphase the 12 daughter chromosomes are approximately cross-shaped, their chromatids being more or less dissociated and divergent from the cen

tromeres. They undergo an equational separation at the second division to form spermatids with the reduced number of chromosomes.

Sitophilus orvzae:

Gonial metaphases of both sexes contain 22 chromosomes (Text fig. 8)-two less than in S. granarius. As the chromosomes come onto the metaphase plate, the presence of 'splits' reveals their bipartite structure and effectively proves, contra Gunson (1945), that 22 is the actual diploid number, as is proved also by counts made at prophase (Text fig. 9). In this species the sex chromosomes are not recognisable, either by heteropycnosis or heteromorphism, throughout the gonial divisions. Pachytene shows that the homologous chromosomes are paired and pre

sent in or about the haploid number in both sexes (see Text fig. 10 for this stage in the male). In that of the female all the elements are non-heteropycnotic; in the male one of the bivalents is composed of a euchromatic and a positively heteropycnotic element joined end-to-end, the latter of which is associated with the nucleolus. This is the sexdetermining pair. In the male, which was alone examined at later stages, all the chromosomes persist in the paired condition through diplotene and diakinesis (Text fig. 11) to metaphase. At this stage (Text figs. 12 to 14) the bivalents are distinctly larger than those of S. granaries (cf. Text figs. 5 to 7). This is a rather rare situation, which

however is known sometimes to differentiate allied species (Tobgy, 1943) in much the same way as allied genera (Kostoff, 1949), and has even been found to distinguish sibling individuals (Thomas, 1936). There appear never to be more than two bivalents, including the largest one,


Text. figs. 8-16. Sitophilus oryzae: 8 & 9, spermatogonial metaphase and prophase,

22 chromosomes some of which are partly separated into their constituent chromatids;

10,'10 autosomal bivalents and the neo-XY pair at pachytene in a male; 11, male dia-

kinesia, 11 bivalents; 12-14, first metaphase in males from Canadian, Australian small,

and Australian large-strain cultures respectively, all with 10 AA+neo-XY; 15, the

lagging of the large bivalent at first anaphase; 16, second spermatocyte metaphases, 11 chromosomes in each.

1952 Cytology of "Sitophilus [Calandrw] oryzae (L.), S. *granarius (L.), etc. 57

that are ring-shaped, whereas in S. granarius at least four autosomal bivalents form rings. Since the spermatogonial chromosomes are obviously in the main metacentric in both species, the failure of the majority to form rings suggests that one of the two arms of most of the chromosomes is largely, if not entirely, heterochromatic, for it has been shown (Smith, 1951 a) that such regions, although completely

paired at pachytene, separate prior to metaphase, that is, they take no part in chiasma formation.

Text figs. 12, 13, and 14 are illustrations of first spermatocyte

metaphases from Canadian, Australian small and Australian large-strain

cultures respectively. The greater size of the chromosomes in Text

fig. 12 is considered to be of no especial significance for other indivi

duals from the same lot have chromosomes like those of the Australian

material. It is assumed that the size difference, unlike that differen

tiating the two species, is a response to a chance variation in treatment

at the time that the slide was prepared.

One of the medium-sized, rod-shaped bivalents frequently appears

at metaphase to be slightly heteromorphic. Hence it is presumably the

sex_??_determiningg pair that at pachytene was seen, to consist of euchro

matin and heterochromatin. Since only two of the three very small

bivalents present in S. granarius are represented in S. oryzae and nei

ther is heteropycnotic or heteromorphic, it is assumed that one or both

of the sex chromosomes as seen in the former species have, in the

latter, become translocated on to a pair of autosomes.

On occasions the largest, ring-shaped bivalent was observed to lag on the equator during anaphase (Text fig. 15), but this was apparently of transitory duration for all second division counts were normal. Neither at early nor late anaphase is there more than the usual tendency

towards dissociation of the component chromatids of the daughter chromosomes. At second metaphase the chromatids are _??_Esociated only at the centromeres (Text fig. 16) providing cross-shaped configurations. This division is normal in all respects, so that spermatids each containing 11 chromosomes are produced.

Sitona lepidus Gyll.:

Spermatogonial divisions were absent from adults of this species. In first spermatocyte metaphases, the ten autosomal bivalents consist of four undoubted rings (at the left in Text fig. 17), two especially large and two smaller, one questionable ring (the eighth from the left),

and five rods; the X and y (at the extreme right) illustrate the conventional 'parachute' form in which these chromosomes are so frequently encountered in beetle spermatogenesis (v. Suomalainen, 1947, figs. 4 to 6 for similar configurations in other Curculionidae). The sex-determining pair in Sitona lepidus is frequently late in congressing, the y


then being closer to the equatorial plate. During pachytene it is posi

tively heteropycnotic and seems always to occur in association with the

nucleolus; during late metaphase the components are faintly negatively

heteropycnotic.

Sitona cylindricollis Fahr.:

This individual was collect

ed on the sandy shores of Lake Michigan, directly south

of Sault Ste. Marie. It has 22 chromosomes which form 11 bivalents at first metaphase

(Text fig. 18). The two largest autosomal bivalents and at least one smaller one are ringshaped. and terminalization is

usually still incomplete in some bivalents at metaphase. The X chro

mosome is somewhat larger than the four smallest autosomes, but the

y is extremely minute, even more so than in S. lepidus, and closely approaches the limit of resolution of the optical system used. The two

are nevertheless clearly associated terminally in the form of an ordinary

parachute. Both meiotic divisions were perfectly orthodox.

Text figs. 17 and 18. Sitona lepidus and Sitona

cylindricollis respectively: first spermatocyte

metaphases (10 AA+Xyp).

Hylobius congener D. S.

and M.:

Males of H. congener. also a curculionid, likewise have a diploid number of 22. For a weevil this species pos

sesses remarkably large chromosomes. They

present so marked a range in size and es

pecially in shape at spermatogonial meta

phase (Text fig. 19) that it proved possible to sort them into homolo

gous pairs with some degree of confidence. Despite its size, smaller than even the smallest arm of an autosome, the y chromosome is

Text figs. 19 and 20. Hylobius congener: 19, spermato

gonial metaphase, 22 chromosomes including the minute mediocentric y and the large mediocentric X chromosome;

20, first spermatocyte metaphase (10 AA+Xyp).

1952 Cytology of Sitophilus [Calandra] oryzae (L.) , S. granarius (L.), etc. 59

without doubt mediocentric, and since there are 11 remaining meta

centrics one must represent the X chromosome and has so been labelled

in Text fig. 19. It is considerably larger than the shortest V-shaped

autosome and somewhat larger than the smallest rod-shaped one.

With one exception the autosomal bivalents in primary spermatocyte metaphases are rod-shaped (Text fig. 20), and the X is now appreciably smaller than the smallest autosome and larger only than the minute y chromosome, with which it associates in the form of a parachute.

Suomalainen (1940 b) found that males of Hylobius abietis have only 21 chromosomes. His species is therefore an XO type, whereas mine has an Xyp sex-determining mechanism. This is not a particularly common occurrence among beetles, although it is known to differentiate allied species in some other families (see Discussion). No very close comparison of the two species is possible, however, since Suomalainen gave no detailed description of his species and failed to provide illustrations of the somatic chromosome complement.

Anthonomus scutellatus Gyll.:

Several adult

males of this minute

species of curculio

nid were, collected

off Aster macro

phyllus L. They

proved to differ

from all weevils examined to date in having 28 chromosomes. Spermatogonia, which were found in only one male, were exceedingly difficult to count owing to the extreme size range of the chromosomes; at least 26 were undoubtedly present. The others contained only spermatocytes of which the later alone were amenable to study. Fourteen bivalents occur at diakinesis and metaphase (Text fig. 21). Usually three but

sometimes only two of the autosomal bivalents are ring-shaped, the remainder being rods. The X chromosome is larger than the two smallest autosomes, which in turn are larger than the y. Together the X and y form a ring in which they are both slightly negatively hetero

pycnotic and closely resemble the sex-determining pair in S. granarius. Ips pini (Say):

TextJfig. 21. Anthonomus scutellatus: first spermatocyte metaphase (13 AA + Xyp).

A single male of this genus in the Scolytidae, newly emerging from a fallen red pine tree, Pinus resinosa Ait., was available for study. It provided an abundance of spermatocyte stages. At first metaphase (Text fig. 22) two or three of the largest of the autosomal bivalents are ring-shaped, the remainder being rods ranging in size from

just smaller than the smallest ring down to extremely minute pairs formed of elements comparable in size with the y chromosome. The


latter, which is large only on this basis, forms a parachute bivalent

with the considerably larger X chromosome. The formula for this species

is therefore 14 AA +Xyp.

Hylurgops pinifex (Fitch):This, the second species in the Scolytidae examined, provided mainly

late spermatocytes, which, however, readily establish its chromosome formula as 14 AA+Xyp. As in the

preceding species, the autosomes cover a considerable size range from relatively large to very small ones

just smaller than the X; the y in this species is again exceptionally minute. Five of the autosomal bivalents are usually associated as rings, and the X and y in

variably form a typical parachute which turns strongly

negatively heteropycnotic by late metaphase (Text fig. 23).

Text figs. 22-24. Scolytidae: first spermatocyte

metaphases in Ips pini, Hylurgops pinifex, and

Dendroctonus engelmanni respectively, each with

the formula 14AA•{Xyp.

Dendroctonus engelmanni Hopk.:

Several adults of this species, the last examined of the scolytidswere shipped from Vernon, B. C., (courtesy Mr. J. Walters), where the host plant is Engelmann spruce, Picea engelmanni Engl. Females proved

to be too old, but males gave spermatocyte stages that show the species to be similar to the two preceding scolytids in having 14 pairs of auto, somes and an Xyp sex-determining system. The size range of the autosomes, however, is less extreme (Text fig. 24) than in Ips and generally smaller than in Hylurgops. It differs further in usually having only one autosomal bivalent in the form of a ring. At pachytene the autosomes have short, strongly heteropycnotic, condensed regions in the neighbourhood of the centromeres, similar to the centric blocks already described in detail in Tribolium confusum Duval (Smith, 1952 a) and other beetles, with long euchromatic segments extending out on one

or both sides.

Euparius marmoreus Oliv.:

This species is a member of the Pla

tystomidae, which, on the basis of larval

differences, has been removed from the

curculionoid superfamily by Boving and

Craighead (1930). At the first meiotic metaphase in males there are 11

Text fig. 25. Platystomidae: first

pro-metaphase (10 AA+Xyp) in the male of Euparius marmoreus.

1952 Cytoiogy of Sitophilus [Calandra] oryzae (L .), S. granarius (L.), etc. 61

bivalents: two large autosomal rings, eight smaller rods, and a typical

parachute-like sex-determining pair. The points of union between the

component chromosomes of all bivalents are terminal at metaphase.

The Xyp bivalent is, as usual, late in congression and by late metaphase

is slightly negatively heteropycnotic. The nucleus illustrated in Text

fig. 25 is at early pro-metaphase, but the bivalents have been drawn

in alignment to facilitate comparison.

Discussion

The validity of earlier findings:It will be at once self-evident from this investigation of the two

species of Sitophilus that the confirmation of Agar's (1911) work on Lepidosiren paradoxes implied in that of Tiegs and Murray and later claimed by Gunson has its basis in erroneous determination and fallacious interpretation. Of this there can not be the slightest doubt: the actual

diploid number in S. oryzae is just twice that reported by Gunson for the male; that of the female (22) is two less than twice the number (12) earlier recorded-the difference arises from an incorrect diagnosis by Gunson of the sex-determining mechanism as being XX: XO instead of XX: XY.

In the light of the present observations all the seeming anomalies

previously reported are readily resolved. Gunson's "...almost twice the diploid number of chromatids..." (actually 22 or 23) in an oogonium

(see her fig. 2) must be accepted as approximately the diploid number of chromosomes; her illustration (fig. 23) of a "Primary spermatocyte nucleus in polar view...", i.e., a metaphase plate, shows "...eleven elements including unpaired chromosome...", in it there is, despite the lateness of the stage, absolutely no evidence of pairing other than a

juxtaposition in pairs of similar sized bodies; the exceptionally clear photomicrograph of an oocyte in polar view at "very late anaphase" (Fig. 8 in Tiegs and Murray's paper) showss eleven daughter chromosomes, but these authors claim that one is a chromosome in process of division, i.e., that there are actually 10 single chromatids and one incompletely divided chromosome; finally, the well-marked constriction in each of the supposed ex-conjugants at metaphase (regarding which Tiegs and Murray state "..11 of the 12 being visible in the section..") are clearly the terminal

points of contact between homologous chromosomes (compare Gunson's fig. 22 with, e. g., Text fig. 14, herein). It is likewise obvious that Inkmann's determination and interpretation are equally incorrect. It must therefore be concluded that the validity of the earlier interpretation as evidence against Darlington's two basic generalizations is groundless5.

5 Upon being apprised of my results, Professor Tiegs rechecked his and Miss

Gunson's preparations and was kind enough to informe me that their original descrip

tions were in error and are now to be interpreted to agree with my own findings.


Agar's observations on Lepidosiren remain in conflict with Darlington's concepts, but this species has never been subjected to reinvestigation

by an independent worker; however, it is significant that they receive no support from Wickbom's (1945) study of Protopterus annectens Owen, another member of the Dipnoi. Similarly, no reinvestigation has been carried out of the actual species that Hogben (1920) examined, but

Dodds (1939) and Speicher (1937) have reported on a member of the Cynipidae and a species of ichneumonid respectively. Dodds states: "The convoluted threads spread throughout the nuclear area and give

rise to ten bivalents, with the homologues of each paired end-to-end

(Fig. 7). Hogben (1919-20), in Neuroterus numismalis, Cynips Kollari and Rhodites rosae, describes the reappearance of the chromatin in the form of twenty separate segments which later pair end-to-end to give ten bivalents. The writer has at no time observed twenty segments at this stage in Neuroterus baccarum. At the earliest stages of the reappearance of the chromatin, the threads are already paired end-toend". Since it is remotely possible that Dodds may have missed the

precise post-diffuse stage that in Hogben's opinion actually precedes the secondary coupling, it cannot be considered that he has effectively proved Hogben wrong, and further investigation is perhaps warranted. Nevertheless, Speicher, working with the obligatory parthenogenetic wasp Ne

meritis canescens (Gray.), presents reasonably conclusive evidence that the chromosomes remain paired through diplotene to metaphase, which casts futher doubt on the validity of Hogben's interpretation.

Comparable cases in other beetles:

One other case in the literature dealing specifically with the cytology of Coleoptera that has been interpreted as an instance of reassociation of ex-conjugant chromosomes has been published by Yosida

(1946). In the testes of two males of a coccinellid, Synonycha grandis Thunberg (2n=18+Xy), he found that some of the cysts showed a number of nuclei in which the chromosomes were completely unpaired. Three types of abnormality are described in the English summary: (a) all the univalents undergo division to produce diploid daughter nuclei;

(b) "At late metaphase the homologous mates come into contact lying side by side. At anaphase, each one of the homologous mates goes to opposite poles (Figs. 12-13). As the result, 10 elements are found in two cells after separation."; and (c) "Separation of the univalent elements in the first division was taken place at random showing irregular distribution of chromosomes (Figs. 18-19)".

Of the three types of abnormality, the first and last are obviously: the expected consequences of complete and partial failure of the chromosomes to pair. The behaviour as described under (b) is, however , most certainly atypical, for in other cysts and in other nuclei in the

1952 Cytology of Sitophilus [Calandra] oryzae (L.) , S. granaries (L.), etc. 63

same cysts, as also in a previous account of meiosis in S. grandis (Yosida, 1944), the meiotic divisions conform to those of Coleoptera in general and of coccinellids in particular (Stevens, 1906 and 1909; Yosida, 1948 and 1951; Smith, unpub.). As, with the exception of the minute y

chromosome, the members of the complement are not markedly different morphologically, the claim that homologous chromosomes associate laterally at late metaphase would be particularly difficult, if not im

possible, to substantiate. Also, if the univalent condition originated through desynapsis rather than through asynapsis, such juxtaposition of homologues would be expected. Finally, since not all nuclei within a cyst were abnormal, the occurrence of daughter nuclei with precisely reduced numbers of chromosomes is of no significance in relation to the reputed lateral association.

Cyto-taxonomic considerations:

Current knowledge of the cytology of curculionids is due primarily to the work of Suomalainen (1940 a and b, 1947 and 1949) and Mikulska

(1949), together with the four contributions made herein. The results of these various studies are brought together in Table I.

With the exception of Hylobius abietis, the species examined by

other workers constitute a rather closely knit taxonomic group in that

they all fall into the subfamily Otiorhynchinae (sense Leng, 1920).

This doubtless contributes in large measure to their general cytological

uniformity, for Strophosomus capitatus, the sole bisexual representative

of the tribe Thylacitini so for examined, alone deviates from the other

wise universal formula 10 AA•{Xy in having an XO sex-determining

mechanism. The remaining curculionids are members of the second

and third subfamilies, the Curculioninae and Calendrinae.

In the Curculioninae, Hylobius abietis resembles H. congener and the two species of Sitona in having 10 pairs of autosomes. The two

species of Hylobius differ from each other, however, in that the latter has an Xyp sex-determining mechanism, whereas the former, like Stro

phosomus capitatus, is devoid of the minute y chromosome. Similarly differentiated species having identical numbers of autosomes are known

in only two other families of beetles: in the Silphidae, Silpha americana has an Xy system (Stevens, 1906) and S. perforata has an XO system

(Yosida, 1951) along with 19 pairs of autosomes; in the Elateridae, the genus Ctenicera (formerly Ludius) includes two complexes of species, one of which has the formula 10AA+Xyp, the other 10AA+XO (Smith, 1952c). The carabid genus Chlaenius likewise includes an Xyp and an XO species, but the former, C. aestivus, has 16 pairs of autosomes (Stevens, 1906) and the latter, C. pallipes, 18 pairs (Yosida, 1951). Considering that more than 300 species of Coleoptera in some 33 families have been investigated to date (Smith, 1952 c), and that failure of pairing with subsequent loss


of a chromosome must be the most common from of meiotic irregularity

leading to numerical change, it is a striking fact that both Xy and XO

sex-determining mechanisms are otherwise known only in the Bruchidae

and Chrysomelidae-two families that are adjacently placed in most

systems of classification. Two explanations suggest themselves to ac

count for this extraordinary distribution: the first, that the genic con

tent of the y, small though it must be, is essential to the survival of

the species; and second, that the mechanical properties of the y are of

considerable importance in ensuring regular segregation of the X chro

mosome. The first, of course, presupposes that the y is otherwise lost, in XO species and not merely translocated into the autosomal comple

ment. Such a shift would, however, introduce the y into the female

line and call for a change in genic balance in both sexes, so that it

would then be plausible to argue that any genes carried on the y are

at least of relatively minor importance.

That the mechanical function of the y is more likely the determining factor is suggested by the greater frequency with which Xy

systems have been replaced by neo-XY ones (Smith, 1949 and 1952 c) for these have arisen in one-third of the families investigated to date. It is suggested also by the fact that once the X has achieved independence in motility (as in XO species), it appears never to have surrendered it by becoming a partner chromosome in a secondary neo-XY system, that is, genera are unknown in which one species is XO and another neo-XY. In neo-XY complexes derived directly from Xyp types, it is again not possible to eliminate the alternative that the y has been retained by translocation onto the autosome that became the neo-Y, except on the basis of the improbability of the simultaneous occurrence of the required number of changes (but see Smith, 1952b).

Two of the three remaining members of the Curculionidae-Cur

culioninae that have been examined, Sitona lepidus and S. cylindricollis,

are what may be termed cytologically normal: they have the formula

10AA+Xyp, which is considered to characterize the Otiorhynchinae.

Using the comparable sizes of their autosomes as a basis for comparison,

they are separable perhaps only owing to the smaller size of the y

chromosome in the latter species.

Anthonomus scutellatus, the remaining member of the Curculioninae

studied, is so isolated both systematically and cytologically that it can

profitably be considered only in relation to the whole complex of species

studied.

The third subfamily of the Curculionidae that I have studied is

the Calendrinae. It was previously mentioned in the section dealing

with materials and methods that no discernible differences could be

found between the chromosomes of the large and small strains of S.

1952 Cytology of Sitophilus [Calandra] oryzae (L.), S. granarius (L.), etc. 65

oryzae obtained from Dr. L. C. Birch, but it should be noted that Birch

(1944) observed a severe limitation to gene exchange between individuals of the two types, which led him to remark: "On the basis of Dobzhansky's (1941, p. 373) criteria for species differences, it would appear that the two forms of C. oryzae could be classed as different species". A preliminary examination of slides made from the two at first suggested that it might be possible to separate them on the basis of the size of the largest bivalent, but more intensive study showed that the difference was referable to unequal rates of contraction during meiosis.

Relative to S. granarius, S. oryzae is obviously a derivative species,

for in it a pair of autosomes has become incorporated into the sex

determining mechanism, that is, it is a neo-XY species. In both species

the heteromorphism of the sex-determining bivalent is indeed only slight:

in S. granarius this is clearly because the X chromosome is inordinately

small; in S. oryzae it is because the incorporation of so small an X into

one of a pair of relatively large autosomes has provided only a compa

ratively negligible resultant increase in its size. In addition, at least

some degree of segmental rearrangement has evidently played a part in

the further phylogenetic divergence of the two species, for, although,

apart from a difference in overall size, the meiotic metaphases of the

two are quite similar, the largest ring in S. oryzae seems to be

proportionately larger than its counterpart in S. granarius.Purely on the basis of comparative chromosome number and mor

phology there is a markedly closer similarity between the only species

of the Platystomidae so far examined and the Curculionidae-Otiorhyn

chinae than there is between Anthonomus scutellatus and its relatives

in the Curculionidae-Curculioninae.

Judging alone from the species now known cytologically it is, of course, not yet possible to suggest whether the chromosomal complement of Anthonomus has been derived from a 10AA+Xyp form, or vice versa. Numerically it is obviously closer to S. granarius, and even the scolytids, than to its neighbouring species in the Curculioninae, and the striking resemblance that the sex chromosomes bear to those of S.

granarius is in conflict with the systematic position to which the latter has been assigned by Roving and Craighead (1931). Whether the cytolo

gical evidence points to a more or less recent origin for Anthonomus can be decided more definitely only if it is first possible to establish what evolutionary trends have been followed by the chromosomes of

the Rhynchophora.

Although the three scolytids have identical chromosome formulae,

they are nevertheless separable through gross morphological differences between their complements. Hylurgops pinifex has considerably bulkier

Cytologia 17, 1952 5


chromosomes than those of Dendroctonus engelmanni and a higher number

of ring bivalents, but the size of the sex chromosomes relative to the

autosomes and the range in size of the latter are comparable in the

two species. On the other hand, the species of Ips, which is in a different

subfamily, shows appreciably greater diversity in the size of its auto

somes and has an X chromosome that exceeds in size even the autosomes

of medium size.

Evolutionary trends:

In two earlier papers (Smith, 1949 and 1950) reasons were presented for adopting the provisional working hypothesis that the primitive formula of the Coleoptera is 9AA+Xyp. The f act of the occurrence in the Carabidae only of species with much higher numbers (v. Makino, 1950)6, parallelling those already reported for Dytiscidae, taken together with the general belief that the Adephaga is a more primitive suborder

than the Polyphaga, might reasonably evoke criticism of such a view. It admittedly opens up the possibility that the chromosome formula of a hypothetical protocoleopteran might perhaps be taken as, say, 18AA

+Xyp, in which the autosomes were all rod-shaped and bore essentially terminal centromeres. Evidence from the Orthoptera (Robertson, 1916) and spiders (Hackman, 1948), two groups in which chromosome evolution has proceeded to some considerable extent by a lowering of number through fusion, might reasonably be put forward as making equally attractive the idea that the Adephaga was early split off from the initial stem, possibly before any extensive reduction in number had been achieved. In agreement with the widespread occurrence of species in it with the formula 9AA+Xyp, it would then be necessary to withdraw the earlier hypothesis in favour of the view that the polyphagan stock had failed to make its appearance until all the autosomes had become V-shaped. The two alternatives thus open to evaluation are therefore:

PROTO-(18AA+Xyp) COLEOPTERAN

PROTOADEPHAGAN (18AA+Xyp)•¨progressive reduction through fusion

reduction through fusion•¨PROTOPOLYPHAGAN

(9AA+Xyp)

PROTO-(9AA+Xyp) COLEOPTERAN

PROTOADEPHAGAN (9AA+Xyp)•¨progressive increase through fragmentation

PROTOPOLYPHAGAN (9AA+Xyp)

Several lines of evidence point to the logical choice of the second alternative: first, the Adephaga has been suggested on the basis of comparative larval morphology (Boving and Craighead, 1931) as being the most recent suborder: second, females of the parthenogenetic species

6 Pending confirmation it appears advisable to ignore the exceptionally low counts reported by Carnoy (1885) judging from his erroneous determination for the hydrophilid Hydrophilus piceus given in the same paper.

1952 Cytology of Sitophilus [Calandra] oryzae (L.). S. granaries (L.), etc. 67

Table I. Summary of chromosome determinations in the Curculionoidea and Scolytidae

based on the classification of Leng (1920)

* Diploid numbers are given for parthenogenetic females. 5*


Micromalthus debilis, the only cytologically known representative of the small, primitive suborder Archostemata, have only ten pairs of chromosomes (Scott, 1936); third, all Cicindelidae (Adephaga) examined to date have either nine or ten pairs of autosomes (Stevens, 1906 and 1909; Goldsmith, 1919; Smith, 1950); fourth, most dytiscids and the only species

in the Gyrinidae for which determinations exist have about 20 pairs of chromosomes (Smith, 1952 c)-this is significant because gyrinids are considered by most authorities to approach the ultimate in adephagan evolution; and, fifth, low numbers, comparable to the 20 of the 9AA+Xyp formula, characterize the Strepsiptera (2n=16; Hughes-Schrader, 1924), the Neuroptera Megaloptera (2n=20 to 26; Itoh, 1933; Naville and de Beaumont, 1936; Klingstedt, 1937), and the Mecoptera (at least as evidenced by Boreus brumalis with 2n=25-Cooper, 1951), the three com

plexes that are believed to be the most closely related to the Coleoptera (Leng, 1920; Crowson, 1950).

The basic formula for the polyphagan stock being taken as 9AA+Xyp and the evolutionary trend one primarily of fragmentation, the phylo

genetic sequence of the groups in Table I might reasonably be: (i) Platystomidae; (ii) Curculionidae-Otiorhynchinae; (iii) Curculionidae-Curculioninae in general; (iv) Curculionidae-Calendrinae (v) Anthonomus; and, finally, (vi) Scolytidae. The various possible means by which the extreme chromosomal diversity evident in the Polyphaga might have come about and the significance of the centric blocks of heterochromatin, as seen herein in Dendroctonus engelmanni, have been dealt with elsewhere

(Smith, 1952 b). As for Boving and Craighead's revision of the families dealt with here, in which they elevate the Platystomidae to superfamily rank, raise the Calendrinae to family status, and merge the

Scolytidae with the remaining Curculionidae and other families to form their superfamily Curculionoidea, the cytological evidence is somewhat ambiguous. It is nevertheless fully evident that the consolidation of the Scolytidae and Curculionidae is, on cytological grounds, diametrically opposed to their erection of the superfamily Platystomoidea.

There are a number of groups of beetles in addition to those already mentioned that provide convincing support for the validity of the conclusion that the two major suborders have a common basic formula. First, there is a general tendency for the chromosome numbers in the Carabidae7 to increase with increasing phylogenetic status; second, many forms which show but little advance over the most primitive polyphagan types show a minimum of cytological change from 9AA+Xy; and, third, many higher chromosome numbers (Yosida, 1949 and 1951) and others

7 Stevens' (1906) determination for Chlaenius pennsylvanicus as 9 AA+Xy is almost

certainly an error, for that of C. aestivus is 16AA+Xy (Stevens, 1909), that of C.

pallipes 18AA+X (Yosida, 1951), and that of C. tricolor 18 AA+X (Smith, unpub.)

1952 Cytology of Sitophilus [Calandra] oryzae (L.), S. granarius (L.). etc. 69

equally lower (Stevens, 1906) exemplify the highly advanced Chrysomelidae. However, some of the lowest numbers at present known are to be found in the Scarabaeidae, Coccinellidae, Buprestidae, and Silphidae, four families which quite certainly would receive very different rank in any system of classification. In particular, along with the especially low number established for the genus Necrophorus (Stevens, 1909; Smith, 1950) there are the particularly high numbers already mentioned for other species in the Silphidae. Possibly in some cases, such for example as Necrophorus and Silpha, this points to the need for a reassessment of prevailing ideas concerning the "closeness" of different genera. Certainly, Roving and Craighead (1931) have carried out a revision of the

Chrysomelidae, as constituted in Long's (1920) synopsis, which has led them to elevate it to superfamily rank. Much of this, however, consists

in raising tribes to subfamily status, subfamilies to family status, and hence the family to superfamily rank. Regardless of the validity of the cytological evidence pro or contra, I would go on record as expressing a strong preference for the retention of old, accepted family names, and as being against innovations that require the use of such expres

sions as "the chrysomeloid complex of families."

Summary

The two calendrid weevils, Sitophilus granarius and S. oryzae, have

the chromosome formulae 11AA+Xyp and 10AA+neo-XY. They there

fore differ from related Curculionidae, bisexual species of which, with

the exception of Anthonomus scutellatus, conform to the formula 10AA

+Xyp or 10AA+XO. Contrary to earlier reports they are nevertheless

meiotically orthodox.

Additional species investigated, with their formulae, are as follows: Platystomidae-Euparius marmoreus=10AA+Xyp; Curculionidae-Sitona lepidus, Sitona cylindricollis, and Hylobius congener=1OAA+Xyp and Anthonomus scutellatus=13AA+Xyp; and Scolytidae-Hylurgops pinifex, Dendroctonus engelmanni, and Ips pini=14AA+Xyp.

These findings are discussed in relation to the taxonomic disposition

of the three families.

References

Agar, W. E. 1911. Quart. J. Micr. Sci. 57: 1-44.Asana, J. J., Makino, S., and Niiyama, H. 1942. Cytologia 12: 187-205.Birch, L. C. 1944. Austral. J. exp. Biol. 22: 271-275.

- 1946. Austral. J. exp. Biol. 24: 123-125.Boving, A. G., and Craighead, F. C. 1931. Ent. Amer. 11: 1-351.Carnoy, J. B. 1885, La Cellule 1: 191-440.Crowson, R. A. 1950. Entom. Monthly Mag. 86: 149-171.Darlington, C. C. 1932. Recent Advances in Cytology. Churchill, London.

- 1937. The Evolution of Genetic Systems. University Press, Cambridge.


Dodds, K. S. 1939. Genetica 21: 177-190.Goldsmith, W. M. 1919. J. Morph. 32: 437-487.Guenin, H. A. 1950. Genetica 25: 157-182.Gunson, M. 1945. Quart. J. Micr. Sci. 85: 107-116.Hackman, W. 1948. Acta Zool. Fenn. 54: 4-101.Hogben, L. T. 1920. Proc. Roy. Soc., B. 91: 268-293.Hughes-Schrader, S. 1924. J. Morph. 39: 157-205.

- 1943. J. Morph. 73: 111-141.Inkmann, F. 1933. Zool. Jahrb., Abt. Anat. 56: 521-558.Itoh, H. 1933. Zool. Mag., Tokyo. 45.Klingstedt, H. 1937. Nature 139: 468.Kostoff, D. 1949. Proc. VIth Int. Cong. Exp. Cyt.: 129-133.Leng, C. W. 1920. Catalogue of Coleoptera of America, north of Mexico. John D.

Sherman Jr., Mount Vernon, N. Y.Makino, S. 1951. An Atlas of the Chromosome Numbers in Animals. The Iowa State

College Press, Ames.Mikulska, I. 1949. Experientia 5: 473.Naville, A., and de Beaumont, J. 1936. Arch. Anat. micr. 32: 271-302.Richard, 0. W. 1944. Trans. Roy. ent. Soc., London 94: 187-200.Robertson, W. R. B. 1916. J. Morph. 27: 179-332.Schrader, F. 1948. Science 107: 155-159.Scott, A. C. 1936. J. Morph. 59: 485-515.Seiler, J. 1947. Chromosoma 3: 88-109.Smith, S. G. 1943. Can. Entom. 75: 21-34.

- 1949. Evolution 3: 344-357.- 1950. Can. Entom. 82: 58-68.- 1951. Genetica 25: 522-524.- 1952a. Chromosoma 4 : 585-610.- 1952b. J. Morph. (in press).- 1952c. Heredity (in press).

Speicher, B. R. 1937. J. Morph. 61: 453-471.Stevens, N. M. 1906. Carneg. Inst. Wash. Pub. No. 36: 33-74.

- 1909. J. Exp. Zool. 6: 101-121.Suomalainen, E. 1940a. Ann. Acad. Sci. Fenn. A 54: 1-144.

- 1940b. Hereditas 26: 51-64.- 1947. Hereditas 33: 425-456.- 1949. Ann. Ent. Fenn. 14: 206-212.

Thomas, P. T. 1936. Nature 138: 402.Tiegs, 0. W., and Murray, F. V. 1938. Quart . J. Micr. Sci. N.S. 80: 159-284. Tobgy, H. A. 1943. J. Gen. 45: 67-111.Wickbom, T. 1945. Hereditas 31: 241-346.Yosida, T. 1944. Jap. J. Genet. 20: 107-115 .

- 1946. Seibutu 1: 218-223.- 1948. Matsumushi 2: 107-109.- 1949. Trans. Sapporo Nat. Hist. Soc . 18** 43-48.- 1951. La Kromosomo 9 (see Makino, 1951) .

the cytology of sitophilus [calandra] oryzae (l.), s ... · the cytology of sitophilus [calandra]...

Documents