Anilocra mediterránea has a diploid chromosome number of twelve, eight of the chromosomes having median, and four subterminal centromeres.

The localization of chiasmata is extreme and is of a previously undescribed type, combining the terminal and centric systems.

The centromeres stain differentially from the non-centric regions of the chromosomes ; they are the only parts stained at anaphase of first meiosis, and at metaphase and anaphase of second meiosis.

Anilocra mediterránea M. Edw. is a fish-louse belonging to the Cymothoid group of Isopods, found abundantly at Naples. In common with other fish-lice, Anilocra is a protandric hermaphrodite, descriptions of its sexual organization having been given by Bullar (1876) and Mayer (1879). For a cytological study of this animal I took individuals while in the predominantly male phase, removing the paired testes and ovaries, and fixed them in La Cour’s 2BD (1937) made up in sea-water. The gonads were fixed for one hour, hardened in 1 per cent, chromic acid for twenty-four hours, washed in running tap-water for half an hour, and then dehydrated and imbedded in wax in the usual manner. Sections were cut at 15/x and stained by the gentian violet-iodine method (see La Cour, 1937).

I examined mitoses in spermatogonia and oogonia and meioses, first and second, in the male germ-cell line. Fixation of the material was satisfactory for general purposes, although not for great detail, the fault lying not in the technique employed but probably in the high degree of hydration of the chromosomes, for other fixatives tried, such as strong Flemming and Bonin in sea-water, all gave indifferent results. In general, mitoses fixed rather better than meioses, especially the prophase stages.

The diploid chromosome number is twelve, eight of the chromosomes having median (M) and four subterminal (8) centromeres. This is a low chromosome number for a crustacean. No particular interest attaches to the mitoses except the variability of the staining reaction, a point to which reference will be made later.

The configurations of the bivalents at first meiotic metaphase show that there is extreme localization of chiasmata in this species. The M-chromosomes regularly form two chiasmata in each arm, one immediately adjacent to the centromere, and one at or near to the end of the arm (see Text-figs. 1 and 2). Four exceptional conditions have been observed: (a) where in one arm a proximal chiasma has failed, (b) where both proximal chiasmata have failed, (c) where in one arm a distal chiasma has failed, and (d) where both distal chiasmata have failed (see Text-figs. 1, 3).

TEXT-FIG. 1.

somewhat diagrammatic. 1. Side view, and 2. polar view, of first meiotic metaphase, normal type. 3. Side view of exceptional configurations of M-chromosomes at first meiotic metaphase.

TEXT-FIG. 1.

somewhat diagrammatic. 1. Side view, and 2. polar view, of first meiotic metaphase, normal type. 3. Side view of exceptional configurations of M-chromosomes at first meiotic metaphase.

TEXT-FIG. 2.

Diagram showing association of M-chromosomes. Xta, chiasmata

TEXT-FIG. 2.

Diagram showing association of M-chromosomes. Xta, chiasmata

TEXT-FIG. 3.

somewhat diagrammatic. 1. Side view of early anaphase of first meiosis. 2. Side view, and 3. polar view, of late anaphase of first meiosis, outlines of unstained chromosomes indicated by dotted lines.

TEXT-FIG. 3.

somewhat diagrammatic. 1. Side view of early anaphase of first meiosis. 2. Side view, and 3. polar view, of late anaphase of first meiosis, outlines of unstained chromosomes indicated by dotted lines.

Out of 365 M-bivalents examined at metaphase, 350 were of the normal type, four of type a, four of type b, five of type c, and two of type d (see Text-fig. 2). The mean chiasma frequency in these bivalents is thus 3·94. The variance from the mean can be calculated from the formula, where V = variance, n = the number of M-bivalents, and x — the chiasma frequency per bivalent. This has a value of 0·087, which represents 0·022 of the mean (see Table I). Values approaching this latter figure have only been found in organisms with extreme localization of chiasmata (Fritillaria ruthenica, Darlington (1936), Meeostethus grossus and Metrioptera brachyptera, Callan, unpub., from data of White (1936)). Haldane (1931) has shown that, where pairing and chiasma-formation are normal, the value of variance approaches the mean, though never reaching this value on account of interference. The variance in this particular case is thus extremely low.

TABLE I.
graphic
graphic

In order to analyse the factors responsible for producing such invariability I have made a study of the proportions in which the various types of association of the M-bivalents occur, with a view to establishing possible correlations. A table is appended (see Table II) showing the number of chiasma failures observed as against the numbers to be expected, assuming absence of correlation between the failures.

TABLE II.

M-bivalents, chromosomes having median centromeres.

M-bivalents, chromosomes having median centromeres.
M-bivalents, chromosomes having median centromeres.

The calculation proceeds as follows :

In the case of the proximal chiasmata, out of 730 possible, 2 ×4+4 = 12 failed.

Thus the chance of a chiasma failing = = 0·0164 = q.

Thus the chance of a chiasma being realized = 1—0–0164 = 0-9836 = p.

The chances of none, one, or two chiasmata failing are, therefore, asp2:2pq:qzrespectively, i.e. as 0-9674:0-03234: 0-0002702.

The corresponding chances for the distal chiasmata failing are 0-9755 : 0-02435 : 0-0001520.

In the table the chances of the various possible combinations of chiasma failure are worked out by multiplication together of their respective components, and then expressed as absolute numbers out of the observed total of 365. The significant squares are those where both proximal chiasmata have failed, the expected number being 0·10 and the observed 4 ; and where both distal chiasmata have failed, the expected number being 0·05 and the observed 2. There is thus evidence of strong positive correlation in the failing of the two proximal chiasmata, and in the failing of the two distal chiasmata. An idea of the relative strengths of these correlations can be obtained from the formula where n0 is the number of bivalents where none, n1 where one, and n2 where two have failed. When there is complete positive correlation this has a value of oo ; when there is no correlation, of 1 ; and when there is complete negative correlation, of 0. For the proximal chiasmata the value is 357 and for the distal chiasmata 115, showing that the former correlation is the stronger.

In the case of the correlation between the failing of the two proximal chiasmata, it would appear that proximity on the chromosome of the two chiasma-forming regions when pairing takes place is the factor of prime importance. When one region has successfully paired there is an increased likelihood for the nearby regions to pair up also, in accordance with the ‘zip’ principle. That there should be, however, a correlation between the failing of the two distal chiasmata cannot be accounted for in this way. One may presume that the two free ends of the zygotene chromosome lie close together, so that if pairing is effected with one of them, the probability is that the other will manage to pair also. Clearly this second correlation indicates that interference plays no part in the restriction of chiasmata distribution down the length of the chromosome: limitation of the initiation of pachytene pairing to the proximal and distal regions and the positions taken up by the zygotene threads within the nucleus are the major factors which secure such invariability and localization of chiasmata as are found in Anilocra.

The S-chromosomes form two chiasmata in the long arm, one proximal and one distal, and no exceptions have been observed ; while in the short arm a chiasma may or may not occur. Often in the same metaphase one of the S-bivalents has the centromeres close together and retained by a chiasma in the short arms, while the other has the centromeres much wider apart and the short arms retracted and not in contact. Such is the position taken up later by the S-bivalents at early anaphase, whether or not a chiasma has occurred in the short arms. In the S-chromosomes there is therefore a corresponding invariability about the association of the long arms such as occurs in the M-chromosomes.

The conclusions drawn from the metaphase configurations cannot be substantiated from a study of diplotene, for the fixation of this stage is very poor. However, the fact that a ‘bouquet’ is formed at pairing stages probably indicates loop formation by the M-chromosomes, an orientation which was inferred from the consideration of the correlation of chiasma failures.

At anaphase the separation of the S-chromosomes precedes that of the M-chromosomes, as is to be expected, since the short arms offer less resistance to separation than do the long arms. The M-chromosomes separate quite regularly and in time with one another, thus providing circumstantial evidence that identical conditions obtain in each one.

The extreme form of chiasma localization found in Anilocra implies a rigidity in the recombination mechanism which is possibly to be correlated with this animal’s parasitic mode of life, and it may be taken to indicate that an evolutionary stand-still has been reached.

A further interesting feature of the meiotic chromosomes is their staining reaction. At full metaphase the bivalents have a bloated appearance and stain lightly ; the centromere, however, is intensely stained and apparently separated from the main body of the chromosome by a non-staining region. At this stage it is triangular in outline, showing tension. During anaphase the stainability of the non-centric regions diminishes, irregular deposits of the stain being laid down on the surface. The stage at which this occurs varies greatly in individual cells. Eventually, soon after the completion of separation, only the centromeres remain stained and six are visible at either pole (see Text-figs. 2 and 3). The outlines of the chromosomes then show only by virtue of a difference in their refractivity from that of the cytoplasm. The centromeres are now rounded off, showing no tension; they are slightly under Ip in size. For comparison, it may be noted that the M-chromosomes at metaphase have a length of 7 μ and width of 2 μ

The stainability of the mitotic chromosomes and of the second meiotic chromosomes is also variable. In the majority of mitoses the anaphase chromosomes remain stained, though sometimes the centromeres only. In the majority of second meioses, even at metaphase, the centromeres only are stained, and in fact the chromosome material remains invisible until late in the maturation of the sperm head. It can thus be seen that there is a progressive tendency through the chromosome cycle, from the somatic divisions to the secondary spermatocyte, for the chromosomes to be unstainable.

Variation in staining reaction given by the same stages of division cannot be the result of differences in fixation through the material, for the unstainable divisions are not distributed through the tissue with any reference to the boundaries of the latter. Furthermore, when there is a certain staining reaction given by one chromosome, the same reaction is given by all the remaining chromosomes in the division. The differential staining reaction is probably the consequence of labile differences in the character of the chromosome surface (gentian violet being a surface stain). This is perhaps dependent on local differences in pH, which would account for the bad fixation of the chromosomes, related as it is to their degree of hydration. It is not the result of a differential distribution of nucleic acid between centromere and non-centric regions. I destained a gentian violet preparation clearly showing the differential staining, and then restained, using the Feulgen technique: no differential staining resulted.

What relation the centromeres, as seen fixed and stained, bear to their size and shape in life is a question of interpretation. It might be argued that what can be stained is really the region of the chromosome near to and including the centromere, not the latter alone. However, its discreteness in stainability and its similar size and shape in all the chromosomes suggests that the artefact is at least an artefact of the centromere itself, and it is of interest to find that in Anilocra there is this clear-cut evidence of a distinction between the centric and non-centric regions of the chromosome.

The fixations were made at Naples during my tenure of the Oxford Biological Scholarship at the Stazione Zoológica for the year 1938/9. The further work was carried out at the John Innes Horticultural Institution, Merton Park. I am much indebted to Dr. Mather for help with the statistical treatment of chiasma-frequency, and to Dr. Darlington for many criticisms and suggestions in the preparation of the argument.

1.
Bullar
,
J. F.
,
1876
. — “
Generative Organs of the Parasitic Isopoda
”,
‘Joum. Anat. Physiol.’
,
11
.
2.
Darlington
,
C. D.
,
1936
. — “
External Mechanics of the Chromosomes
”,
‘Proc. Roy. Soc.’, B
,
121
.
3.
Haldane
,
J. B. S.
,
1931
. — “
Cytological Basis of Genetical Interference
”,
‘Cytologia’
,
3
.
4.
La Cour
,
L
.,
1937
. — “
Improvements in Plant Cytological Technique
”,
‘Botanical Review’
,
5
.
5.
Mayer
,
P.
,
1879
. — “
Carcinologische Mittheihmgen, 6. Hermaphroditismus bei einigen Isopoden
”,
‘Mitt. Zool. Stat. Neapel.’
,
1
.
6.
White
,
M. J. D.
,
1936
. — “
Chiasma localisation in Mecostethus grossus, and Metrioptera brachyptera (Orthoptera)
”,
‘Zeits. Zellf. u. mikr. Anat.’
,
24
.