By splitting the bramble-stumps in the course of July, we perceive also that the Three-pronged Osmia, notwithstanding her narrow gallery, follows the same practice as Latreille's Osmia, with a difference. She does not build a party-wall, which the diameter of the cylinder would not permit; she confines herself to putting up a frail circular pad of green putty, as though to limit, before any attempt at harvesting, the space to be occupied by the Bee-bread, whose depth could not be calculated afterwards if the insect did not first mark out its confines. Can there really be an act of measuring? That would be superlatively clever. Let us consult the Three-horned Osmia in her glass tubes.

The Osmia is working at her big partition, with her body outside the cell which she is preparing. From time to time, with a pellet of mortar in her mandibles, she goes in and touches the previous ceiling with her forehead, while the tip of her abdomen quivers and feels the pad in course of construction. One might well say that she is using the length of her body as a measure, in order to fix the next ceiling at the proper distance. Then she resumes her work. Perhaps the measure was not correctly taken; perhaps her memory, a few seconds old, has already become muddled. The Bee once more ceases laying her plaster and again goes and touches the front wall with her forehead and the back wall with the tip of her abdomen. Looking at that body trembling with eagerness, extended to its full length to touch the two ends of the room, how can we fail to grasp the architect's grave problem? The Osmia is measuring; and her measure is her body. Has she quite done, this time? Oh dear no! Ten times, twenty times, at every moment, for the least particle of mortar which she lays, she repeats her mensuration, never being quite certain that her trowel is going just where it should.

Meanwhile, amid these frequent interruptions, the work progresses and the partition gains in width. The worker is bent into a hook, with her mandibles on the inner surface of the wall and the tip of her abdomen on the outer surface. The soft masonry stands between the two points of purchase. The insect thus forms a sort of rolling-press, in which the mud wall is flattened and shaped. The mandibles tap and furnish mortar; the end of the abdomen also pats and gives brisk trowel-touches. This anal extremity is a builder's tool; I see it facing the mandibles on the other side of the partition, kneading and smoothing it all over, flattening the little lump of clay. It is a singular implement, which I should never have expected to see used for this purpose. It takes an insect to conceive such an original idea, to do mason's work with its behind! During this curious performance, the only function of the legs is to keep the worker steady by spreading out and clinging to the walls of the tunnel.

The partition with the hole in it is finished. Let us go back to the measuring of which the Osmia was so lavish. What a magnificent argument in favour of the reasoning-power of animals! To find geometry, the surveyor's art, in an Osmia's tiny brain! An insect that begins by taking the measurements of the room to be constructed, just as any master-builder might do! Why, it's splendid, it's enough to cover with confusion those horrible sceptics who persist in refusing to admit the animal's 'continuous little flashes of atoms of reason!'

O common-sense, veil your face! It is with this gibberish about continuous flashes of atoms of reason that men pretend to build up science to-day! Very well, my masters; the magnificent argument with which I am supplying you lacks but one little detail, the merest trifle: truth! Not that I have not seen and plainly seen all that I am relating; but measuring has nothing to do with the case. And I can prove it by facts.

If, in order to see the Osmia's nest as a whole, we split a reed lengthwise, taking care not to disturb its contents; or, better still, if we select for examination the string of cells built in a glass tube, we are forthwith struck by one detail, namely, the uneven distances between the partitions, which are placed almost at right angles to the axis of the cylinder. It is these distances which fix the size of the chambers, which, with a similar base, have different heights and consequently unequal holding-capacities. The bottom partitions, the oldest, are farther apart; those of the front part, near the orifice, are closer together. Moreover, the provisions are plentiful in the loftier cells, whereas they are niggardly and reduced to one-half or even one-third in the cells of lesser height.

Here are a few examples of these inequalities. A glass tube with a diameter of 12 millimetres (.468 inch.—Translator's Note.), inside measurement, contains ten cells. The five lower ones, beginning with the bottom-most, have as the respective distances between their partitions, in millimetres:

11, 12, 16, 13, 11. (.429,.468,.624,.507,.429 inch.—Translator's Note.)

The five upper ones measure between their partitions:

7, 7, 5, 6, 7. (.273,.273,.195,.234,.273 inch.—Translator's Note.)

A reed-stump 11 millimetres (.429 inch.—Translator's Note.) across the inside contains fifteen cells; and the respective distances between the partitions of those cells, starting from the bottom, are:

13, 12, 12, 9, 9, 11, 8, 8, 7, 7, 7, 6, 6, 6, 7. (.507,.468,.468, .351,.351,.429,.312,.312,.273,.273,.273,.234,.234,.234, .273 inch.—Translator's Note.)

When the diameter of the tunnel is less, the partitions can be still further apart, though they retain the general characteristic of being closer to one another the nearer they are to the orifice. A reed of five millimetres (.195 inch.—Translator's Note.) in diameter, gives me the following distances, always starting from the bottom:

22, 22, 20, 20, 12, 14. (.858,.858,.78,.78,.468,.546 inch.—Translator's Note.)

Another, of 9 millimetres (.351 inch.—Translator's Note.), gives me:

15, 14, 11, 10, 10, 9, 10. (.585,.546,.429,.39,.39,.351,.39 inch.—Translator's Note.)

A glass tube of 8 millimetres (.312 inch.—Translator's Note.) yields:

15, 14, 20, 10, 10, 10. (.585,.546,.78,.39,.39,.39 inch.—Translator's Note.).

I could fill pages and pages with such figures, if I cared to print all my notes. Do they prove that the Osmia is a geometrician, employing a strict measure based on the length of her body? Certainly not, because many of those figures exceed the length of the insect; because sometimes a higher number follows suddenly upon a lower; because the same string contains a figure of one value and another figure of but half that value. They prove only one thing: the marked tendency of the insect to shorten the distance between the party-walls as the work proceeds. We shall see later that the large cells are destined for the females and the small ones for the males.

Is there not at least a measuring adapted to each sex? Again, not so; for in the first series, where the females are housed, instead of the interval of 11 millimetres, which occurs at the beginning and the end, we find, in the middle of the series, an interval of 16 millimetres, while in the second series, reserved for the males, instead of the interval of 7 millimetres at the beginning and the end, we have an interval of 5 millimetres in the middle. It is the same with the other series, each of which shows a striking discrepancy in its figures. If the Osmia really studied the dimensions of her chambers and measured them with the compasses of her body, how could she, with her delicate mechanism, fail to notice mistakes of 5 millimetres, almost half her own length?

Besides, all idea of geometry vanishes if we consider the work in a tube of moderate width. Here, the Osmia does not fix the front partition in advance; she does not even lay its foundation. Without any boundary-pad, with no guiding mark for the capacity of the cell, she busies herself straightway with the provisioning. When the heap of Bee-bread is judged sufficient, that is, I imagine, when her tired body tells her that she has done enough harvesting, she closes up the chamber. In this case, there is no measuring; and yet the capacity of the cell and the quantity of the victuals fulfil the regular requirements of one or the other sex.

Then what does the Osmia do when she repeatedly stops to touch the front partition with her forehead and the back partition, the one in the course of building, with the tip of her abdomen? I have no idea what she does or what she has in view. I leave the interpretation of this performance to others, more venturesome than I. Plenty of theories are based on equally shaky foundations. Blow on them and they sink into the quagmire of oblivion.

The laying is finished, or perhaps the cylinder is full. A final partition closes the last cell. A rampart is now built, at the orifice of the tube itself, to forbid the ill-disposed all access to the home. This is a thick plug, a massy work of fortification, whereon the Osmia spends enough mortar to partition off any number of cells. A whole day is not too long for making this barricade, especially in view of the minute finishing-touches, when the Osmia fills up with putty every chink through which the least atom could slip. The mason completing a wall smooths his plaster and brings it to a fine surface while it is still wet; the Osmia does the same, or almost. With little taps of the mandibles and a continual shaking of her head, a sign of her zest for the work, she smooths and polishes the surface of the lid for hours at a time. After such pains, what foe could visit the dwelling?

And yet there is one, an Anthrax, A. sinuata (Cf. "The Life of the Fly": chapters 2 and 4.—Translator's Note.), who will come later on, in the height of summer, and succeed, invisible bit of thread that she is, in making her way to the grub through the thickness of the door and the web of the cocoon. In many cells, mischief of another kind has already been done. During the progress of the works, an impudent Midge, one of the Tachina-flies, who feeds her family on the victuals amassed by the Bee, hovers in front of the galleries. Does she penetrate to the cells and lay her eggs there in the mother's absence? I could never catch the sneak in the act. Does she, like that other Tachina who ravages cells stocked with game (The cells of the Hunting Wasps.—Translator's Note.), nimbly deposit her eggs on the Osmia's harvest at the moment when the Bee is going indoors? It is possible, though I cannot say for certain. The fact remains that we soon see the Midge's grub-worms swarming around the larva, the daughter of the house. There are ten, fifteen, twenty or more of them gnawing with their pointed mouths at the common dish and turning the food into a heap of fine, orange-coloured vermicelli. The Bee's grub dies of starvation. It is life, life in all its ferocity even in these tiny creatures. What an expenditure of ardent labour, of delicate cares, of wise precautions, to arrive at...what? Her offspring sucked and drained dry by the hateful Anthrax; her family sweated and starved by the infernal Tachina.

The victuals consist mostly of yellow flour. In the centre of the heap, a little honey is disgorged, which turns the pollen-dust into a firm, reddish paste. On this paste the egg is laid, not flat, but upright, with the fore-end free and the hind-end lightly held and fixed in the plastic mass. When hatched, the young grub, kept in its place by its rear-end, need only bend its neck a little to find the honey-soaked paste under its mouth. When it grows stronger, it will release itself from its support and eat up the surrounding flour.

All this is touching, in its maternal logic. For the new-born, dainty bread-and-honey; for the adolescent, dry bread. In cases where the provisions are all of a kind, these delicate precautions are superfluous. The victuals of the Anthophorae and the Chalicodomae consist of flowing honey, the same throughout. The egg is then laid at full length on the surface, without any particular arrangement, thus compelling the new-born grub to take its first mouthfuls at random. This has no drawback, as the food is of the same quality throughout. But, with the Osmia's provisions—dry powder on the edges, jam in the centre—the grub would be in danger if its first meal were not regulated in advance. To begin with pollen not seasoned with honey would be fatal to its stomach. Having no choice of its mouthfuls because of its immobility and being obliged to feed on the spot where it was hatched, the young grub must needs be born on the central mass, where it has only to bend its head a little way in order to find what its delicate stomach calls for. The place of the egg, therefore, fixed upright by its base in the middle of the red jam, is most judiciously chosen. What a contrast between this exquisite maternal forethought and the horrible destruction by the Anthrax and the Midge!

The egg is rather large for the size of the Osmia. It is cylindrical, slightly curved, rounded at both ends and transparent. It soon becomes cloudy, while remaining diaphanous at each extremity. Fine lines, hardly perceptible to the most penetrating lens, show themselves in transverse circles. These are the first signs of segmentation. A contraction appears in the front hyaline part, marking the head. An extremely thin opaque thread runs down either side. This is the cord of tracheae communicating between one breathing-hole and another. At last, the segments show distinctly, with their lateral pads. The grub is born.

At first, one would think that there was no hatching in the proper sense of the word—that is to say, no bursting and casting of a wrapper. The most minute attention is necessary to show that appearances are deceptive and that actually a fine membrane is thrown off from front to back. This infinitesimal shred is the shell of the egg.

The grub is born. Fixed by its base, it curves into an arc and bends its head, until now held erect, down to the red mass. The meal begins. Soon a yellow cord occupying the front two-thirds of the body proclaims that the digestive apparatus is swelling out with food. For a fortnight, consume your provender in peace, my child; then spin your cocoon: you are now safe from the Tachina! Shall you be safe from the Anthrax' sucker later on? Alack!



CHAPTER 3. THE DISTRIBUTION OF THE SEXES.

Does the insect know beforehand the sex of the egg which it is about to lay? When examining the stock of food in the cells just now, we began to suspect that it does, for each little heap of provisions is carefully proportioned to the needs at one time of a male and at another of a female. What we have to do is to turn this suspicion into a certainty demonstrated by experiment. And first let us find out how the sexes are arranged.

It is not possible to ascertain the chronological order of a laying, except by going to suitably-chosen species. Digging up the burrows of Cerceris-, Bembex- or Philanthus-wasps will never tell us that this grub has taken precedence of that in point of time nor enable us to decide whether one cocoon in a colony belongs to the same family as another. To compile a register of births is absolutely impossible here. Fortunately there are a few species in which we do not find this difficulty: these are the Bees who keep to one gallery and build their cells in storeys. Among the number are the different inhabitants of the bramble-stumps, notably the Three-pronged Osmiae, who form an excellent subject for observation, partly because they are of imposing-size—bigger than any other bramble-dwellers in my neighbourhood—partly because they are so plentiful.

Let us briefly recall the Osmia's habits. Amid the tangle of a hedge, a bramble-stalk is selected, still standing, but a mere withered stump. In this the insect digs a more or less deep tunnel, an easy piece of work owing to the abundance of soft pith. Provisions are heaped up right at the bottom of the tunnel and an egg is laid on the surface of the food: that is the first-born of the family. At a height of some twelve millimetres (About half an inch.—Translator's Note.), a partition is fixed, formed of bramble saw-dust and of a green paste obtained by masticating particles of the leaves of some plant that has not yet been identified. This gives a second storey, which in its turn receives provisions and an egg, the second in order of primogeniture. And so it goes on, storey by storey, until the cylinder is full. Then a thick plug of the same green material of which the partitions are formed closes the home and keeps out marauders.

In this common cradle, the chronological order of births is perfectly clear. The first-born of the family is at the bottom of the series; the last-born is at the top, near the closed door. The others follow from bottom to top in the same order in which they followed in point of time. The laying is numbered automatically; each cocoon tells us its respective age by the place which it occupies.

To know the sexes, we must wait for the month of June. But it would be unwise to postpone our investigations until that period. Osmia-nests are not so common that we can hope to pick one up each time that we go out with that object; besides, if we wait for the hatching-period before examining the brambles, it may happen that the order has been disturbed through some insects' having tried to make their escape as soon as possible after bursting their cocoons; it may happen that the male Osmiae, who are more forward than the females, are already gone. I therefore set to work a long time beforehand and devote my leisure in winter to these investigations.

The bramble-sticks are split and the cocoons taken out one by one and methodically transferred to glass tubes, of approximately the same diameter as the native cylinder. These cocoons are arranged one on top of the other in exactly the same order that they occupied in the bramble; they are separated from one another by a cotton plug, an insuperable obstacle to the future insect. There is thus no fear that the contents of the cells may become mixed or transposed; and I am saved the trouble of keeping a laborious watch. Each insect can hatch at its own time, in my presence or not: I am sure of always finding it in its place, in its proper order, held fast fore and aft by the cotton barrier. A cork or sorghum-pith partition would not fulfil the same purpose: the insect would perforate it and the register of births would be muddled by changes of position. Any reader wishing to undertake similar investigations will excuse these practical details, which may facilitate his work.

We do not often come upon complete series, comprising the whole laying, from the first-born to the youngest. As a rule, we find part of a laying, in which the number of cocoons varies greatly, sometimes falling as low as two, or even one. The mother has not deemed it advisable to confide her whole family to a single bramble-stump; in order to make the exit less toilsome, or else for reasons which escape me, she has left the first home and elected to make a second home, perhaps a third or more.

We also find series with breaks in them. Sometimes, in cells distributed at random, the egg has not developed and the provisions have remained untouched, but mildewed; sometimes, the larva has died before spinning its cocoon, or after spinning it. Lastly, there are parasites, such as the Unarmed Zonitis (Zonitis mutica, one of the Oil-beetles.—Translator's Note.) and the Spotted Sapyga (A Digger-wasp.—Translator's Note.), who interrupt the series by substituting themselves for the original occupant. All these disturbing factors make it necessary to examine a large number of nests of the Three-pronged Osmia, if we would obtain a definite result.

I have been studying the bramble-dwellers for seven or eight years and I could not say how many strings of cocoons have passed through my hands. During a recent winter, in view particularly of the distribution of the sexes, I collected some forty of this Osmia's nests, transferred their contents into glass tubes and made a careful summary of the sexes. I give some of my results. The figures start in their order from the bottom of the tunnel dug in the bramble and proceed upwards to the orifice. The figure 1 therefore denotes the first-born of the series, the oldest in date; the highest figure denotes the last-born. The letter M, placed under the corresponding figure, represents the male and the letter F the female sex.

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 F F M F M F M M F F F F M F M

This is the longest series that I have ever been able to procure. It is also complete, inasmuch as it comprises the entire laying of the Osmia. My statement requires explaining, otherwise it would seem impossible to know whether a mother whose acts one has not watched, nay more, whom one has never seen, has or has not finished laying her eggs. The bramble-stump under consideration leaves a free space of nearly four inches above the continuous string of cocoons. Beyond it, at the actual orifice, is the terminal stopper, the thick plug which closes the entrance to the gallery. In this empty portion of the tunnel there is ample accommodation for numerous cocoons. The fact that the mother has not made use of it proves that her ovaries were exhausted; for it is exceedingly unlikely that she has abandoned first-rate lodgings to go laboriously digging a new gallery elsewhere and there continue her laying.

You may say that, if the unoccupied space marks the end of the laying, nothing tells us that the beginning is actually at the bottom of the cul-de-sac, at the other end of the tunnel. You may also say that the laying is done in shifts, separated by intervals of rest. The space left empty in the channel would mean that one of these shifts was finished and not that there were no more eggs ripe for hatching. In answer to these very plausible explanations, I will say that, the sum of my observations—and they have been extremely numerous—is that the total number of eggs laid not only by the Osmiae but by a host of other Bees fluctuates round about fifteen.

Besides, when we consider that the active life of these insects lasts hardly a month; when we remember that this period of activity is disturbed by dark, rainy or very windy days, during which all work is suspended; when lastly we ascertain, as I have done ad nauseam in the case of the Three-horned Osmia, the time required for building and victualling a cell, it becomes obvious that the total laying must be kept within narrow bounds and that the mother has no time to lose if she wishes to get fifteen cells satisfactorily built in three or four weeks interrupted by compulsory rests. I shall give some facts later which will dispel your doubts, if any remain.

I assume, therefore, that a number of eggs bordering on fifteen represents the entire family of an Osmia, as it does of many other Bees.

Let us consult some other complete series. Here are two:

1 2 3 4 5 6 7 8 9 10 11 12 13 F F M F M F M F F F F M F F M F F F M F F M F M

In both cases, the laying is taken as complete, for the same reasons as above.

We will end with some series that appear to me incomplete, in view of the small number of cells and the absence of any free space above the pile of cocoons:

1 2 3 4 5 6 7 8 M M F M M M M M M M F M F M M M F M F F M M M M M F M F F F F M M M F M

These examples are more than sufficient. It is quite evident that the distribution of the sexes is not governed by any rule. All that I can say on consulting the whole of my notes, which contain a good many instances of complete layings—most of them, unfortunately, spoilt through gaps caused by parasites, the death of the larva, the failure of the egg to hatch and other accidents—all that I can say in general is that the complete series begins with females and nearly always ends with males. The incomplete series can teach us nothing in this respect, for they are only fragments starting we know not whence; and it is impossible to tell whether they should be ascribed to the beginning, to the end, or to an intermediate period of the laying. To sum up: in the laying of the Three-pronged Osmia, no order governs the succession of the sexes; only, the series has a marked tendency to begin with females and to finish with males.

The brambles, in my district, harbour two other Osmiae, both of much smaller size: O. detrita, PEREZ, and O. parvula, DUF. The first is very common, the second very rare; and until now I have found only one of her nests, placed above a nest of O. detrita, in the same bramble. Here, instead of the lack of order in the distribution of the sexes which we find with O. tridentata, we have an order remarkable for consistency and simplicity. I have before me the list of the series of O. detrita collected last winter. Here are some of them:

1. A series of twelve: seven females, beginning with the bottom of the tunnel, and then five males.

2. A series of nine: three females first, then six males.

3. A series of eight: five females followed by three males.

4. A series of eight: seven females followed by one male.

5. A series of eight: one female followed by seven males.

6. A series of seven: six females followed by one male.

The first series might very well be complete. The second and fifth appear to be the end of layings, of which the beginning has taken place elsewhere, in another bramble-stump. The males predominate and finish off the series. Nos. 3, 4 and 6, on the other hand, look like the beginnings of layings: the females predominate and are at the head of the series. Even if these interpretations should be open to doubt, one result at least is certain: with O. detrita, the laying is divided into two groups, with no intermingling of the sexes; the first group laid yields nothing but females, the second, or more recent, yields nothing but males.

What was only a sort of attempt with the Three-pronged Osmia—who, it is true, begins with females and ends with males, but muddles up the order and mixes the two sexes anyhow between the extreme points—becomes a regular law with her kinswoman. The mother occupies herself at the start with the stronger sex, the more necessary, the better-gifted, the female sex, to which she devotes the first flush of her laying and the fullness of her vigour; later, when she is perhaps already at the end of her strength, she bestows what remains of her maternal solicitude upon the weaker sex, the less-gifted, almost negligible male sex.

O. parvula, of whom I unfortunately possess but one series, repeats what the previous witness has just shown us. This series, one of nine cocoons, comprises five females followed by four males, without any mixing of the sexes.

Next to these disgorgers of honey and gleaners of pollen-dust, it would be well to consult other Hymenoptera, Wasps who devote themselves to the chase and pile their cells one after the other, in a row, showing the relative age of the cocoons. The brambles house several of these: Solenius vagus, who stores up Flies; Psen atratus, who provides her grubs with a heap of Plant-lice; Trypoxylon figulus, who feeds them with Spiders.

Solenius vagus digs her gallery in a bramble-stick that is lopped short, but still fresh and green. The house of this Fly-huntress, therefore, suffers from damp, as the sap enters, especially on the lower floors. This seems to me rather insanitary. To avoid the humidity, or for other reasons which escape me, the Solenius does not dig very far into her bramble-stump and consequently can stack but a small number of cells in it. A series of five cocoons gives me first four females and then one male; another series, also of five, contains first three females, with two males following. These are the most complete that I have for the moment.

I reckoned on the Black Psen, or Psen atratus, whose series are pretty long; it is a pity that they are nearly always greatly interfered with by a parasite called Ephialtes mediator. (Cf. "The Life of the Fly": chapter 2.—Translator's Note.) I obtained only three series free from gaps: one of eight cocoons, comprising only females; one of six, likewise consisting wholly of females; lastly, one of eight, formed exclusively of males. These instances seem to show that the Psen arranges her laying in a succession of females and a succession of males; but they tell us nothing of the relative order of the two series.

From the Spider-huntress, Trypoxylon figulus, I learnt nothing decisive. She appeared to me to rove about from one bramble to the next, utilizing galleries which she has not dug herself. Not troubling to be economical with a lodging which it has cost her nothing to acquire, she carelessly builds a few partitions at very unequal heights, stuffs three or four compartments with Spiders and passes on to another bramble-stump, with no reason, so far as I know, for abandoning the first. Her cells, therefore, occur in series that are too short to give us any useful information.

This is all that the bramble-dwellers have to tell us; I have enumerated the list of the principal ones in my district. We will now look into some other Bees who arrange their cocoons in single files: the Megachiles (Cf. Chapter 8 of the present volume.—Translator's Note.), who cut disks out of leaves and fashion the disks into thimble-shaped receptacles; the Anthidia (Cf. Chapters 9 and 10 of the present volume.—Translator's Note.), who weave their honey-wallets out of cotton-wool and arrange their cells one after the other in some cylindrical gallery. In most cases, the home is the produce of neither the one nor the other. A tunnel in the upright, earthy banks, the old work of some Anthophora, is the usual dwelling. There is no great depth to these retreats; and all my searches, zealously prosecuted during a number of winters, procured me only series containing a small number of cocoons, four or five at most, often one alone. And, what is quite as serious, nearly all these series are spoilt by parasites and allow me to draw no well-founded deductions.

I remembered finding, at rare intervals, nests of both the Anthidium and the Megachile in the hollows of cut reeds. I thereupon installed some hives of a new kind on the sunniest walls of my enclosure. They consisted of stumps of the great reed of the south, open at one end, closed at the other by the natural knot and gathered into a sort of enormous pan-pipe, such as Polyphemus might have employed. The invitation was accepted: Osmiae, Anthidia and Megachiles came in fairly large numbers, especially the first, to benefit by the queer installation.

In this way I obtained some magnificent series of Anthidia and Megachiles, running up to a dozen. There was a melancholy side to this success. All my series, with not one exception, were ravaged by parasites. Those of the Megachile (M. sericans, FONSCOL), who fashions her goblets with robinia-, holm-, and terebinth-leaves, were inhabited by Coelioxys octodentata (A Parasitic Bee.—Translator's Note.); those of the Anthidium (A. florentinum, LATR.) were occupied by a Leucopsis. Both kinds were swarming with a colony of pigmy parasites whose name I have not yet been able to discover. In short, my pan-pipe hives, though very useful to me from other points of view, taught me nothing about the order of the sexes among the Leaf-cutters and the cotton-weavers.

I was more fortunate with three Osmiae (O. tricornis, LATR., O. cornuta, LATR., and O. Latreillii, SPIN.), all of whom gave me splendid results, with reed-stumps arranged either against the walls of my garden, as I have just said, or near their customary abode, the huge nests of the Mason-bee of the Sheds. One of them, the Three-horned Osmia, did better still: as I have described, she built her nests in my study, as plentifully as I could wish, using reeds, glass tubes and other retreats of my selecting for her galleries.

We will consult this last, who has furnished me with documents beyond my fondest hopes, and begin by asking her of how many eggs her average laying consists. Of the whole heap of colonized tubes in my study, or else out of doors, in the hurdle-reeds and the pan-pipe appliances, the best-filled contains fifteen cells, with a free space above the series, a space showing that the laying is ended, for, if the mother had any more eggs available, she would have lodged them in the room which she leaves unoccupied. This string of fifteen appears to be rare; it was the only one that I found. My attempts at indoor rearing, pursued during two years with glass tubes or reeds, taught me that the Three-horned Osmia is not much addicted to long series. As though to decrease the difficulties of the coming deliverance, she prefers short galleries, in which only a part of the laying is stacked. We must then follow the same mother in her migration from one dwelling to the next if we would obtain a complete census of her family. A spot of colour, dropped on the Bee's thorax with a paint-brush while she is absorbed in closing up the mouth of the tunnel, enables us to recognize the Osmia in her various homes.

In this way, the swarm that resided in my study furnished me, in the first year, with an average of twelve cells. Next year, the summer appeared to be more favourable and the average became rather higher, reaching fifteen. The most numerous laying performed under my eyes, not in a tube, but in a succession of Snail-shells, reached the figure of twenty-six. On the other hand, layings of between eight and ten are not uncommon. Lastly, taking all my records together, the result is that the family of the Osmia fluctuates round about fifteen in number.

I have already spoken of the great differences in size apparent in the cells of one and the same series. The partitions, at first widely spaced, draw gradually nearer to one another as they come closer to the aperture, which implies roomy cells at the back and narrow cells in front. The contents of these compartments are no less uneven between one portion and another of the string. Without any exception known to me, the large cells, those with which the series starts, have more abundant provisions than the straitened cells with which the series ends. The heap of honey and pollen in the first is twice or even thrice as large as that in the second. In the last cells, the most recent in date, the victuals are but a pinch of pollen, so niggardly in amount that we wonder what will become of the larva with that meagre ration.

One would think that the Osmia, when nearing the end of the laying, attaches no importance to her last-born, to whom she doles out space and food so sparingly. The first-born receive the benefit of her early enthusiasm: theirs is the well-spread table, theirs the spacious apartments. The work has begun to pall by the time that the last eggs are laid; and the last-comers have to put up with a scurvy portion of food and a tiny corner.

The difference shows itself in another way after the cocoons are spun. The large cells, those at the back, receive the bulky cocoons; the small ones, those in front, have cocoons only a half or a third as big. Before opening them and ascertaining the sex of the Osmia inside, let us wait for the transformation into the perfect insect, which will take place towards the end of summer. If impatience gets the better of us, we can open them at the end of July or in August. The insect is then in the nymphal stage; and it is easy, under this form, to distinguish the two sexes by the length of the antennae, which are larger in the males, and by the glassy protuberances on the forehead, the sign of the future armour of the females. Well, the small cocoons, those in the narrow front cells, with their scanty store of provisions, all belong to males; the big cocoons, those in the spacious and well-stocked cells at the back, all belong to females.

The conclusion is definite: the laying of the Three-horned Osmia consists of two distinct groups, first a group of females and then a group of males.

With my pan-pipe apparatus displayed on the walls of my enclosure and with old hurdle-reeds left lying flat out of doors, I obtained the Horned Osmia in fair quantities. I persuaded Latreille's Osmia to build her nest in reeds, which she did with a zeal which I was far from expecting. All that I had to do was to lay some reed-stumps horizontally within her reach, in the immediate neighbourhood of her usual haunts, namely, the nests of the Mason-bee of the Sheds. Lastly, I succeeded without difficulty in making her build her nests in the privacy of my study, with glass tubes for a house. The result surpassed my hopes.

With both these Osmiae, the division of the gallery is the same as with the Three-horned Osmia. At the back are large cells with plentiful provisions and widely-spaced partitions; in front, small cells, with scanty provisions and partitions close together. Also, the larger cells supplied me with big cocoons and females; the smaller cells gave me little cocoons and males. The conclusion therefore is exactly the same in the case of all three Osmiae.

Before dismissing the Osmiae, let us devote a moment to their cocoons, a comparison of which, in the matter of bulk, will furnish us with fairly accurate evidence as to the relative size of the two sexes, for the thing contained, the perfect insect, is evidently proportionate to the silken wrapper in which it is enclosed. These cocoons are oval-shaped and may be regarded as ellipsoids formed by a revolution around the major axis. The volume of one of these solids is expressed in the following formula:

4 / 3 x pi x a x (b squared),

in which 2a is the major axis and 2b the minor axis.

Now, the average dimensions of the cocoons of the Three-horned Osmia are as follows:

2a = 13 mm. (.507 inch.—Translator's Note.), 2b = 7 mm. (.273 inch.—Translator's Note.) in the females;

2a = 9 mm. (.351 inch.—Translator's Note.), 2b = 5 mm. (.195 inch.—Translator's Note.) in the males.

The ratio therefore between 13 x 7 x 7 = 637 and 9 x 5 x 5 = 225 will be more or less the ratio between the sizes of the two sexes. This ratio is somewhere between 2 to 1 and 3 to 1. The females therefore are two or three times larger than the males, a proportion already suggested by a comparison of the mass of provisions, estimated simply by the eye.

The Horned Osmia gives us the following average dimensions:

2a = 15 mm. (.585 inch.—Translator's Note.), 2b = 9 mm. (.351 inch.—Translator's Note.) in the females;

2a = 12 mm. (.468 inch.—Translator's Note.), 2b = 7 mm. (.273 inch.—Translator's Note.) in the males.

Once again, the ratio between 15 x 9 x 9 = 1215 and 12 x 7 x 7 = 588 lies between 2 to 1 and 3 to 1.

Besides the Bees who arrange their laying in a row, I have consulted others whose cells are grouped in a way that makes it possible to ascertain the relative order of the two sexes, though not quite so precisely. One of these is the Mason-bee of the Walls. I need not describe again her dome-shaped nest, built on a pebble, which is now so well-known to us. (Cf. "The Mason-bees": chapter 1.—Translator's Note.)

Each mother chooses her stone and works on it in solitude. She is an ungracious landowner and guards her site jealously, driving away any Mason who even looks as though she might alight on it. The inhabitants of the same nest are therefore always brothers and sisters; they are the family of one mother.

Moreover, if the stone presents a large enough surface—a condition easily fulfilled—the Mason-bee has no reason to leave the support on which she began her laying and go in search of another whereon to deposit the rest of her eggs. She is too thrifty of her time and of her mortar to involve herself in such expenditure except for grave reasons. Consequently, each nest, at least when it is new, when the Bee herself has laid the first foundations, contains the entire laying. It is a different thing when an old nest is restored and made into a place for depositing the eggs. I shall come back later to such houses.

A newly-built nest then, with rare exceptions, contains the entire laying of one female. Count the cells and we shall have the total list of the family. Their maximum number fluctuates round about fifteen. The most luxuriant series will occasionally reach as many as eighteen, though these are very scarce.

When the surface of the stone is regular all around the site of the first cell, when the mason can add to her building with the same facility in every direction, it is obvious that the groups of cells, when finished, will have the oldest in the central portion and the more recent in the surrounding portion. Because of this juxtaposition of the cells, which serve partly as a wall to those which come next, it is possible to form some estimate of the chronological order of the cells in the Chalicodoma's nest and thus to discover the sequence of the two sexes.

In winter, by which time the Bee has long been in the perfect state, I collect Chalicodoma-nests, removing them bodily from their support with a few smart sideward taps of the hammer on the pebbles. At the base of the mortar dome the cells are wide agape and display their contents. I take the cocoon from its box, open it and take note of the sex of the insect enclosed.

I should probably be accused of exaggeration if I mentioned the total number of the nests which I have gathered and the cells which I have inspected by this method during the last six or seven years. I will content myself with saying that the harvest of a single morning sometimes consisted of as many as sixty nests of the Mason-bee. I had to have help in carrying home my spoils, even though the nests were removed from their stones on the spot.

From the enormous number of nests which I have examined, I am able to state that, when the cluster is regular, the female cells occupy the centre and the male cells the edges. Where the irregularity of the pebble has prevented an even distribution around the initial point, the same rule has been observed. A male cell is never surrounded on every side by female cells: either it occupies the edges of the nest, or else it adjoins, at least on some sides, other male cells, of which the last form part of the exterior of the cluster. As the surrounding cells are obviously of a later date than the inner cells, it follows that the Mason-bee acts like the Osmiae: she begins her laying with females and ends it with males, each of the sexes forming a series of its own, independent of the other.

Some further circumstances add their testimony to that of the surrounded and surrounding cells. When the pebble projects sharply and forms a sort of dihedral angle, one of whose faces is more or less vertical and the other horizontal, this angle is a favourite site with the Mason, who thus finds greater stability for her edifice in the support given her by the double plane. These sites appear to me to be in great request with the Chalicodoma, considering the number of nests which I find thus doubly supported. In nests of this kind, all the cells, as usual, have their foundations fixed to the horizontal surface; but the first row, the row of cells first built, stands with its back against the vertical surface.

Well, these older cells, which occupy the actual edge of the dihedral angle, are always female, with the exception of those at either end of the row, which, as they belong to the outside, may be male cells. In front of this first row come others. The female cells occupy the middle portion and the male the ends. Finally, the last row, closing in the remainder, contains only male cells. The progress of the work is very visible here: the Mason has begun by attending to the central group of female cells, the first row of which occupies the dihedral angle, and has finished her task by distributing the male cells round the outside.

If the perpendicular face of the dihedral angle be high enough, it sometimes happens that a second row of cells is placed above the first row backing on to that plane; a third row occurs less often. The nest is then one of several storeys. The lower storeys, the older, contain only females; the upper, the more recent storey, contains none but males. It goes without saying that the surface layer, even of the lower storeys, can contain males without invalidating the rule, for this layer may always be looked upon as the Chalicodoma's last work.

Everything therefore contributes to show that, in the Mason-bee, the females take the lead in the order of primogeniture. Theirs is the central and best-protected part of the clay fortress; the outer part, that most exposed to the inclemencies of the weather and to accidents, is for the males.

The males' cells do not differ from the females' only by being placed at the outside of the cluster; they differ also in their capacity, which is much smaller. To estimate the respective capacities of the two sorts of cells, I go to work as follows: I fill the empty cell with very fine sand and pour this sand back into a glass tube measuring 5 millimetres (.195 inch.—Translator's Note.) in diameter. From the height of the column of sand we can estimate the comparative capacity of the two kinds of cells. I will take one at random among my numerous examples of cells thus gauged.

It comprises thirteen cells and occupies a dihedral angle. The female cells give me the following figures, in millimetres, as the height of the columns of sand:

40, 44, 43, 48, 48, 46, 47 (1.56, 1.71, 1.67, 1.87, 1.87, 1.79, 1.83 inches.—Translator's Note.),

averaging 45. (1.75 inches.—Translator's Note.)

The male cells give me:

32, 35, 28, 30, 30, 31 (1.24, 1.36, 1.09, 1.17, 1.17, 1.21 inches.—Translator's Note.),

averaging 31. (1.21 inches.—Translator's Note.)

The ratio of the capacity of the cells for the two sexes is therefore roughly a ratio of 4 to 3. The actual contents of the cell being proportionate to its capacity, the above ratio must also be more or less the ratio of provisions and sizes between females and males. These figures will assist us presently to tell whether an old cell, occupied for a second or third time, belonged originally to a female or a male.

The Chalicodoma of the Sheds cannot give us any information on this matter. She builds under the same eaves, in excessively populous colonies; and it is impossible to follow the labours of any single Mason, whose cells, distributed here and there, are soon covered up with the work of her neighbours. All is muddle and confusion in the individual output of the swarming throng.

I have not watched the work of the Chalicodoma of the Shrubs with close enough attention to be able to state definitely that this Bee is a solitary builder. Her nest is a ball of clay hanging from a bough. Sometimes, this nest is the size of a large walnut and then appears to be the work of one alone; sometimes, it is the size of a man's fist, in which case I have no doubt that it is the work of several. Those bulky nests, comprising more than fifty cells, can tell us nothing exact, as a number of workers must certainly have collaborated to produce them.

The walnut-sized nests are more trustworthy, for everything seems to indicate that they were built by a single Bee. Here females are found in the centre of the group and males at the circumference, in somewhat smaller cells, thus repeating what the Mason-bee of the Pebbles has told us.

One clear and simple rule stands out from this collection of facts. Apart from the strange exception of the Three-pronged Osmia, who mixes the sexes without any order, the Bees whom I studied and probably a crowd of others produce first a continuous series of females and then a continuous series of males, the latter with less provisions and smaller cells. This distribution of the sexes agrees with what we have long known of the Hive-bee, who begins her laying with a long sequence of workers, or sterile females, and ends it with a long sequence of males. The analogy continues down to the capacity of the cells and the quantities of provisions. The real females, the Queen-bees, have wax cells incomparably more spacious than the cells of the males and receive a much larger amount of food. Everything therefore demonstrates that we are here in the presence of a general rule.

But does this rule express the whole truth? Is there nothing beyond a laying in two series? Are the Osmiae, the Chalicodomae and the rest of them fatally bound by this distribution of the sexes into two distinct groups, the male group following upon the female group, without any mixing of the two? Is the mother absolutely powerless to make a change in this arrangement, should circumstances require it?

The Three-pronged Osmia already shows us that the problem is far from being solved. In the same bramble-stump, the two sexes occur very irregularly, as though at random. Why this mixture in the series of cocoons of a Bee closely related to the Horned Osmia and the Three-horned Osmia, who stack theirs methodically by separate sexes in the hollow of a reed? What the Bee of the brambles does cannot her kinswomen of the reeds do too? Nothing, so far as I know, can explain this difference in a physiological act of primary importance. The three Bees belong to the same genus; they resemble one another in general outline, internal structure and habits; and, with this close similarity, we suddenly find a strange dissimilarity.

There is just one thing that might possibly arouse a suspicion of the cause of this irregularity in the Three-pronged Osmia's laying. If I open a bramble-stump in the winter to examine the Osmia's nest, I find it impossible, in the vast majority of cases, to distinguish positively between a female and a male cocoon: the difference in size is so small. The cells, moreover, have the same capacity: the diameter of the cylinder is the same throughout and the partitions are almost always the same distance apart. If I open it in July, the victualling-period, it is impossible for me to distinguish between the provisions destined for the males and those destined for the females. The measurement of the column of honey gives practically the same depth in all the cells. We find an equal quantity of space and food for both sexes.

This result makes us foresee what a direct examination of the two sexes in the adult form tells us. The male does not differ materially from the female in respect of size. If he is a trifle smaller, it is scarcely noticeable, whereas, in the Horned Osmia and the Three-horned Osmia, the male is only half or a third the size of the female, as we have seen from the respective bulk of their cocoons. In the Mason-bee of the Walls there is also a difference in size, though less pronounced.

The Three-pronged Osmia has not therefore to trouble about adjusting the dimensions of the dwelling and the quantity of the food to the sex of the egg which she is about to lay; the measure is the same from one end of the series to the other. It does not matter if the sexes alternate without order: one and all will find what they need, whatever their position in the row. The two other Osmiae, with their great disparity in size between the two sexes, have to be careful about the twofold consideration of board and lodging. And that, I think, is why they begin with spacious cells and generous rations for the homes of the females and end with narrow, scantily-provisioned cells, the homes of the males. With this sequence, sharply defined for the two sexes, there is less fear of mistakes which might give to one what belongs to another. If this is not the explanation of the facts, I see no other.

The more I thought about this curious question, the more probable it appeared to me that the irregular series of the Three-pronged Osmia and the regular series of the other Osmiae, of the Chalicodomae and of the Bees in general were all traceable to a common law. It seemed to me that the arrangement in a succession first of females and then of males did not account for everything. There must be something more. And I was right: that arrangement in series is only a tiny fraction of the reality, which is remarkable in a very different way. This is what I am going to prove by experiment.



CHAPTER 4. THE MOTHER DECIDES THE SEX OF THE EGG.

I will begin with the Mason-bee of the Pebbles. (This is the same insect as the Mason-bee of the Walls. Cf. "The Mason-bees": passim.—Translator's Note.) The old nests are often used, when they are in good enough repair. Early in the season the mothers quarrel fiercely over them; and, when one of the Bees has taken possession of the coveted dome, she drives any stranger away from it. The old house is far from being a ruin, only it is perforated with as many holes as it once had occupants. The work of restoration is no great matter. The heap of earth due to the destruction of the lid by the outgoing tenant is taken out of the cell and flung away at a distance, atom by atom. The remnants of the cocoon are also thrown away, but not always, for the delicate silken wrapper sometimes adheres closely to the masonry.

The victualling of the renovated cell is now begun. Next comes the laying; and lastly the orifice is sealed with a mortar plug. A second cell is utilized in the same way, followed by a third and so on, one after the other, as long as any remain unoccupied and the mother's ovaries are not exhausted. Finally, the dome receives, mainly over the apertures already plugged, a coat of plaster which makes the nest look like new. If she has not finished her laying, the mother goes in search of other old nests to complete it. Perhaps she does not decide to found a new establishment except when she can find no second-hand dwellings, which mean a great economy of time and labour. In short, among the countless number of nests which I have collected, I find many more ancient than recent ones.

How shall we distinguish one from the other? The outward aspect tells you nothing, owing to the great care taken by the Mason to restore the surface of the old dwelling equal to new. To resist the rigours of the winter, this surface must be impregnable. The mother knows that and therefore repairs the dome. Inside, it is another matter: the old nest stands revealed at once. There are cells whose provisions, at least a year old, are intact, but dried up or musty, because the egg has never developed. There are others containing a dead larva, reduced by time to a blackened, curled-up cylinder. There are some whence the perfect insect was never able to issue: the Chalicodoma wore herself out in trying to pierce the ceiling of her chamber; her strength failed her and she perished in the attempt. Others again and very many are occupied by ravagers, Leucopses (Cf. "The Mason-bees": chapter 11.—Translator's Note.) and Anthrax-flies, who will come out a good deal later, in July. Altogether, the house is far from having every room vacant; there are nearly always a considerable number occupied either by parasites that were still in the egg-stage at the time when the Mason-bee was at work or by damaged provisions, dried grubs or Chalicodomae in the perfect state who have died without being able to effect their deliverance.

Should all the rooms be available, a rare occurrence, there still remains a method of distinguishing between an ancient nest and a recent one. The cocoon, as I have said, adheres pretty closely to the walls; and the mother does not always take away this remnant, either because she is unable to do so, or because she considers the removal unnecessary. Thus the base of the new cocoon is set in the bottom of the old cocoon. This double wrapper points very clearly to two generations, two separate years. I have even found as many as three cocoons fitting one into another at their bases. Consequently, the nests of the Mason-bee of the Pebbles are able to do duty for three years, if not more. Eventually they become utter ruins, abandoned to the Spiders and to various smaller Bees or Wasps, who take up their quarters in the crumbling rooms.

As we see, an old nest is hardly ever capable of containing the Mason-bee's entire laying, which calls for some fifteen apartments. The number of rooms at her disposal is most unequal, but always very small. It is saying much when there are enough to receive about half the laying. Four or five cells, sometimes two or even one: that is what the Mason usually finds in a nest that is not her own work. This large reduction is explained when we remember the numerous parasites that live upon the unfortunate Bee.

Now, how are the sexes distributed in those layings which are necessarily broken up between one old nest and another? They are distributed in such a way as utterly to upset the idea of an invariable succession first of females and then of males, the idea which occurs to us on examining the new nests. If this rule were a constant one, we should be bound to find in the old domes at one time only females, at another only males, according as the laying was at its first or at its second stage. The simultaneous presence of the two sexes would then correspond with the transition period between one stage and the next and should be very unusual. On the contrary, it is very common; and, however few cells there may be, we always find both females and males in the old nests, on the sole condition that the compartments have the regulation holding-capacity, a large capacity for the females, a lesser for the males, as we have seen.

The old male cells can be recognized by their position on the outer edges and by their capacity, measuring on an average the same as a column of sand 31 millimetres high in a glass tube 5 millimetres wide. (1.21 x.195 inches.—Translator's Note.) These cells contain males of the second or third generation and none but males. In the old female cells, those in the middle, whose capacity is measured by a similar column of sand 45 millimetres high (1.75 inches.—Translator's Note.), are females and none but females.

This presence of both sexes at a time, even when there are but two cells free, one spacious and the other small, proves in the plainest fashion that the regular distribution observed in the complete nests of recent production is here replaced by an irregular distribution, harmonizing with the number and holding-capacity of the chambers to be stocked. The Mason-bee has before her, let me suppose, only five vacant cells: two larger and three smaller. The total space at her disposal would do for about a third of the laying. Well, in the two large cells, she puts females; in the three small cells, she puts males.

As we find the same sort of thing in all the old nests, we must needs admit that the mother knows the sex of the egg which she is going to lay, because that egg is placed in a cell of the proper capacity. We can go further and admit that the mother alters the order of succession of the sexes at her pleasure, because her layings, between one old nest and another, are broken up into small groups of males and females according to the exigencies of space in the actual nest which she happens to be occupying.

Just now, in the new nest, we saw the Mason-bee arranging her total laying into series first of females and next of males; and here she is, mistress of an old nest of which she has not the power to alter the arrangement, breaking up her laying into sections comprising both sexes just as required by the conditions imposed upon her. She therefore decides the sex of the egg at will, for, without this prerogative, she could not, in the chambers of the nest which she owes to chance, deposit unerringly the sex for which those chambers were originally built; and this happens however small the number of chambers to be filled.

When the nest is new, I think I see a reason why the Mason-bee should seriate her laying into females and then males. Her nest is a half-sphere. That of the Mason-bee of the Shrubs is very nearly a sphere. Of all shapes, the spherical shape is the strongest. Now these two nests require an exceptional power of resistance. Without protection of any kind, they have to brave the weather, one on its pebble, the other on its bough. Their spherical configuration is therefore very practical.

The nest of the Mason-bee of the Walls consists of a cluster of upright cells backing against one another. For the whole to take a spherical form, the height of the chambers must diminish from the centre of the dome to the circumference. Their elevation is the sine of the meridian arc starting from the plane of the pebble. Therefore, if they are to have any solidity, there must be large cells in the middle and small cells at the edges. And, as the work begins with the central chambers and ends with those on the circumference, the laying of the females, destined for the large cells, must precede that of the males, destined for the small cells. So the females come first and the males at the finish.

This is all very well when the mother herself founds the dwelling, when she lays the first rows of bricks. But, when she is in the presence of an old nest, of which she is quite unable to alter the general arrangement, how is she to make use of the few vacant rooms, the large and the small alike, if the sex of the egg be already irrevocably fixed? She can only do so by abandoning the arrangement in two consecutive rows and accommodating her laying to the varied exigencies of the home. Either she finds it impossible to make an economical use of the old nest, a theory refuted by the evidence, or else she determines at will the sex of the egg which she is about to lay.

The Osmiae themselves will furnish the most conclusive evidence on the latter point. We have seen that these Bees are not generally miners, who themselves dig out the foundation of their cells. They make use of the old structures of others, or else of natural retreats, such as hollow stems, the spirals of empty shells and various hiding-places in walls, clay or wood. Their work is confined to repairs to the house, such as partitions and covers. There are plenty of these retreats; and the insect would always find first-class ones if it thought of going any distance to look for them. But the Osmia is a stay-at-home: she returns to her birth-place and clings to it with a patience extremely difficult to exhaust. It is here, in this little familiar corner, that she prefers to settle her progeny. But then the apartments are few in number and of all shapes and sizes. There are long and short ones, spacious ones and narrow. Short of expatriating herself, a Spartan course, she has to use them all, from first to last, for she has no choice. Guided by these considerations, I embarked on the experiments which I will now describe.

I have said how my study, on two separate occasions, became a populous hive, in which the Three-horned Osmia built her nests in the various appliances which I had prepared for her. Among these appliances, tubes, either of glass or reed, predominated. There were tubes of all lengths and widths. In the long tubes, entire or almost entire layings, with a series of females followed by a series of males, were deposited. As I have already referred to this result, I will not discuss it again. The short tubes were sufficiently varied in length to lodge one or other portion of the total laying. Basing my calculations on the respective lengths of the cocoons of the two sexes, on the thickness of the partitions and the final lid, I shortened some of these to the exact dimensions required for two cocoons only, of different sexes.

Well, these short tubes, whether of glass or reed, were seized upon as eagerly as the long tubes. Moreover, they yielded this splendid result: their contents, only a part of the total laying, always began with female and ended with male cocoons. This order was invariable; what varied was the number of cells in the long tubes and the proportion between the two sorts of cocoons, sometimes males predominating and sometimes females.

The experiment is of paramount importance; and it will perhaps make the result clearer if I quote one instance from among a multitude of similar cases. I give the preference to this particular instance because of the rather exceptional fertility of the laying. An Osmia marked on the thorax is watched, day by day, from the commencement to the end of her work. From the 1st to the 10th of May, she occupies a glass tube in which she lodges seven females followed by a male, which ends the series. From the 10th to the 17th of May, she colonizes a second tube, in which she lodges first three females and then three males. From the 17th to the 25th of May, a third tube, with three females and then two males. On the 26th of May, a fourth tube, which she abandons, probably because of its excessive width, after laying one female in it. Lastly, from the 26th to the 30th of May, a fifth tube, which she colonizes with two females and three males. Total: twenty-five Osmiae, including seventeen females and eight males. And it will not be superfluous to observe that these unfinished series do not in any way correspond with periods separated by intervals of rest. The laying is continuous, in so far as the variable condition of the atmosphere allows. As soon as one tube is full and closed, another is occupied by the Osmia without delay.

The tubes reduced to the exact length of two cells fulfilled my expectation in the great majority of cases: the lower cell was occupied by a female and the upper by a male. There were a few exceptions. More discerning than I in her estimate of what was strictly necessary, better-versed in the economy of space, the Osmia had found a way of lodging two females where I had only seen room for one female and a male.

This experiment speaks volumes. When confronted with tubes too small to receive all her family, she is in the same plight as the Mason-bee in the presence of an old nest. She thereupon acts exactly as the Chalicodoma does. She breaks up her laying, divides it into series as short as the room at her disposal demands; and each series begins with females and ends with males. This breaking up, on the one hand, into sections in all of which both sexes are represented and the division, on the other hand, of the entire laying into just two groups, one female, the other male, when the length of the tube permits, surely provide us with ample evidence of the insect's power to regulate the sex of the egg according to the exigencies of space.

And besides the exigencies of space one might perhaps venture to add those connected with the earlier development of the males. These burst their cocoons a couple of weeks or more before the females; they are the first who hasten to the sweets of the almond-tree. In order to release themselves and emerge into the glad sunlight without disturbing the string of cocoons wherein their sisters are still sleeping, they must occupy the upper end of the row; and this, no doubt, is the reason that makes the Osmia end each of her broken layings with males. Being next to the door, these impatient ones will leave the home without upsetting the shells that are slower in hatching.

I experimented on Latreille's Osmia, using short and even very short stumps of reed. All that I had to do was to lay them just beside the nests of the Mason-bee of the Sheds, nests beloved by this particular Osmia. Old, disused hurdles supplied me with reeds inhabited from end to end by the Horned Osmia. In both cases I obtained the same results and the same conclusions as with the Three-horned Osmia.

I return to the latter, nidifying under my eyes in some old nests of the Mason-bee of the Walls, which I had placed within her reach, mixed up with the tubes. Outside my study, I had never yet seen the Three-horned Osmia adopt that domicile. This may be due to the fact that these nests are isolated one by one in the fields; and the Osmia, who loves to feel herself surrounded by her kin and to work in plenty of company, refuses them because of this isolation. But on my table, finding them close to the tubes in which the others are working, she adopts them without hesitation.

The chambers presented by those old nests are more or less spacious according to the thickness of the coat of mortar which the Chalicodoma has laid over the assembled chambers. To leave her cell, the Mason-bee has to perforate not only the plug, the lid built at the mouth of the cell, but also the thick plaster wherewith the dome is strengthened at the end of the work. The perforation results in a vestibule which gives access to the chamber itself. It is this vestibule which is sometimes longer and sometimes shorter, whereas the corresponding chamber is of almost constant dimensions, in the case of the same sex, of course.

We will first consider the short vestibule, at the most large enough to receive the plug with which the Osmia will close up the lodging. There is then nothing at her disposal except the cell proper, a spacious apartment in which one of the Osmia's females will find ample accommodation, for she is much smaller than the original occupant of the chamber, no matter the sex; but there is not room for two cocoons at a time, especially in view of the space taken up by the intervening partition. Well, in those large, well-built chambers, formerly the homes of Chalicodomae, the Osmia settles females and none but females.

Let us now consider the long vestibule. Here, a partition is constructed, encroaching slightly on the cell proper, and the residence is divided into two unequal storeys, a large room below, housing a female, and a narrow cabin above, containing a male.

When the length of the vestibule permits, allowing for the space required by the outer stopper, a third storey is built, smaller than the second; and another male is lodged in this cramped corner. In this way the old nest of the Mason-bee of the Pebbles is colonized, cell after cell, by a single mother.

The Osmia, as we see, is very frugal of the lodging that has fallen to her share; she makes the best possible use of it, giving to the females the spacious chambers of the Mason-bee and to the males the narrow vestibules, subdivided into storeys when this is feasible. Economy of space is the chief consideration, since her stay-at-home tastes do not allow her to indulge in distant quests. She has to employ the site which chance places at her disposal just as it is, now for a male and now for a female. Here we see displayed, more clearly than ever, her power of deciding the sex of the egg, in order to adapt it judiciously to the conditions of the house-room available.

I had offered at the same time to the Osmiae in my study some old nests of the Mason-bee of the Shrubs, which are clay spheroids with cylindrical cavities in them. These cavities are formed, as in the old nests of the Mason-bee of the Pebbles, of the cell properly so-called and of the exit-way which the perfect insect cut through the outer coating at the time of its deliverance. Their diameter is about seven millimetres (.273 inch.—Translator's Note.); their depth at the centre of the heap is 23 millimetres (.897 inch.—Translator's Note.); and at the edge averages 14 millimetres (.546 inch.—Translator's Note.)

The deep central cells receive only the females of the Osmia; sometimes even the two sexes together, with a partition in the middle, the female occupying the lower and the male the upper storey. True, in such cases economy of space is strained to the utmost, the apartments provided by the Mason-bee of the Shrubs being very small as it is, despite their entrance-halls. Lastly, the deeper cavities on the circumference are allotted to females and the shallower to males.

I will add that a single mother peoples each nest and also that she proceeds from cell to cell without troubling to ascertain the depth. She goes from the centre to the edges, from the edges to the centre, from a deep cavity to a shallow cavity and vice versa, which she would not do if the sexes were to follow upon each other in a settled order. For greater certainty, I numbered the cells of one nest as each of them was closed. On opening them later, I was able to see that the sexes were not subjected to a chronological arrangement. Females were succeeded by males and these by females without its being possible for me to make out any regular sequence. Only—and this is the essential point—the deep cavities were allotted to the females and the shallow ones to the males.

We know that the Three-horned Osmia prefers to haunt the habitations of the Bees who nidify in populous colonies, such as the Mason-bee of the Sheds and the Hairy-footed Anthophora. Exercising the very greatest care, I broke up some great lumps of earth removed from the banks inhabited by the Anthophora and sent to me from Carpentras by my dear friend and pupil M. Devillario. I examined them conscientiously in the quiet of my study. I found the Osmia's cocoons arranged in short series, in very irregular passages, the original work of which is due to the Anthophora. Touched up afterwards, made larger or smaller, lengthened or shortened, intersected with a network of crossings by the numerous generations that had succeeded one another in the same city, they formed an inextricable labyrinth.

Sometimes these corridors did not communicate with any adjoining apartment; sometimes they gave access to the spacious chamber of the Anthophora, which could be recognized, in spite of its age, by its oval shape and its coating of glazed stucco. In the latter case, the bottom cell, which once constituted, by itself, the chamber of the Anthophora, was always occupied by a female Osmia. Beyond it, in the narrow corridor, a male was lodged, not seldom two, or even three. Of course, clay partitions, the work of the Osmia, separated the different inhabitants, each of whom had his own storey, his own closed cell.

When the accommodation consisted of no more than a simple cylinder, with no state-bedroom at the end of it—a bedroom always reserved for a female—the contents varied with the diameter of the cylinder. The series, of which the longest were series of four, included, with a wider diameter, first one or two females, then one or two males. It also happened, though rarely, that the series was reversed, that is to say, it began with males and ended with females. Lastly, there were a good many isolated cocoons, of one sex or the other. When the cocoon was alone and occupied the Anthophora's cell, it invariably belonged to a female.

I have observed the same thing in the nests of the Mason-bee of the Sheds, but not so easily. The series are shorter here, because the Mason-bee does not bore galleries but builds cell upon cell. The work of the whole swarm thus forms a stratum of cells that grows thicker from year to year. The corridors occupied by the Osmia are the holes which the Mason-bee dug in order to reach daylight from the deep layers. In these short series, both sexes are usually present; and, if the Mason-bee's chamber is at the end of the passage, it is inhabited by a female Osmia.

We come back to what the short tubes and the old nests of the Mason-bee of the Pebbles have already taught us. The Osmia who, in tubes of sufficient length, divides her whole laying into a continuous sequence of females and a continuous sequence of males, now breaks it up into short series in which both sexes are present. She adapts her sectional layings to the exigencies of a chance lodging; she always places a female in the sumptuous chamber which the Mason-bee or the Anthophora occupied originally.

Facts even more striking are supplied by the old nests of the Masked Anthophora (A. personata, ILLIG.), old nests which I have seen utilized by the Horned Osmia and the Three-horned Osmia at the same time. Less frequently, the same nests serve for Latreille's Osmia. Let us first describe the Masked Anthophora's nests.

In a steep bank of sandy clay, we find a set of round, wide-open holes. There are generally only a few of them, each about half an inch in diameter. They are the entrance-doors leading to the Anthophora's abode, doors always left open, even after the building is finished. Each of them gives access to a short passage, sometimes straight, sometimes winding, nearly horizontal, polished with minute care and varnished with a sort of white glaze. It looks as if it had received a thin coat of whitewash. On the inner surface of this passage, in the thickness of the earthy bank, spacious oval niches have been excavated, communicating with the corridor by means of a narrow bottle-neck, which is closed, when the work is done, with a substantial mortar stopper. The Anthophora polishes the outside of this stopper so well, smooths its surface so perfectly, bringing it to the same level as that of the passage, is so careful to give it the white tint of the rest of the wall that, when the job is finished, it becomes absolutely impossible to distinguish the entrance-door corresponding with each cell.

The cell is an oval cavity dug in the earthy mass. The wall has the same polish, the same chalky whiteness as the general passage. But the Anthophora does not limit herself to digging oval niches: to make her work more solid, she pours over the walls of the chamber a salivary liquid which not only whitens and varnishes but also penetrates to a depth of some millimetres into the sandy earth, which it turns into a hard cement. A similar precaution is taken with the passage; and therefore the whole is a solid piece of work capable of remaining in excellent condition for years.

Moreover, thanks to the wall hardened by the salivary fluid, the structure can be removed from its matrix by chipping it carefully away. We thus obtain, at least in fragments, a serpentine tube from which hangs a single or double row of oval nodules that look like large grapes drawn out lengthwise. Each of these nodules is a cell, the entrance to which, carefully hidden, opens into the tube or passage. When she wishes to leave her cell, in the spring, the Anthophora destroys the mortar disk that closes the jar and thus reaches the general corridor, which is quite open to the outer air. The abandoned nest provides a series of pear-shaped cavities, of which the distended part is the old cell and the contracted part the exit-neck, rid of its stopper.

These pear-shaped hollows form splendid lodgings, impregnable strongholds, in which the Osmiae find a safe and commodious retreat for their families. The Horned Osmia and the Three-horned Osmia establish themselves there at the same time. Although it is a little too large for her, Latrielle's Osmia also appears very well satisfied with it.

I have examined some forty of the superb cells utilized by each of the first two. The great majority are divided into two storeys by means of a transversal partition. The lower storey includes the larger portion of the Anthophora's cell; the upper storey includes the rest of the cell and a little of the bottle-neck that surmounts it. The two-roomed dwelling is closed, in the passage, by a shapeless, bulky mass of dried mud. What a clumsy artist the Osmia is, compared with the Anthophora! Against the exquisite work of the Anthophora, partition and plug strike a note as hideously incongruous as a lump of dirt on polished marble.

The two apartments thus obtained are of a very unequal capacity, which at once strikes the observer. I measured them with my five-millimetre tube. On an average, the bottom one is represented by a column of sand 50 millimetres deep (1.95 inches.—Translator's Note.) and the top one by a column of 15 millimetres (.585 inch.—Translator's Note.). The holding-capacity of the one is therefore about three times as large as that of the other. The cocoons enclosed present the same disparity. The bottom one is big, the top one small. Lastly, the lower one belongs to a female Osmia and the upper to a male Osmia.

Occasionally the length of the bottle-neck allows of a fresh arrangement and the cavity is divided into three storeys. The bottom one, which is always the most spacious, contains a female; the two above, both smaller than the first and one smaller than the other, contain males.

Let us keep to the first case, which is always the most frequent. The Osmia is in the presence of one of these pear-shaped hollows. It is a find that must be employed to the best advantage: a prize of this sort is rare and falls only to fortune's favourites. To lodge two females in it at once is impossible; there is not sufficient room. To lodge two males in it would be undue generosity to a sex that is entitled to but the smallest consideration. Besides, the two sexes must be represented in almost equal numbers. The Osmia decides upon one female, whose portion shall be the better room, the lower one, which is larger, better-protected and more nicely polished, and one male, whose portion shall be the upper storey, a cramped attic, uneven and rugged in the part which encroaches on the bottle-neck. This decision is proved by numerous undeniable facts. Both Osmiae therefore can choose the sex of the egg about to be laid, seeing that they are now breaking up the laying into groups of two, a female and a male, as required by the conditions of the lodging.

I have only once found Latreille's Osmia established in the nest of the Masked Anthophora. She had occupied but a small number of cells, because the others were not free, being inhabited by the Anthophora. The cells in question were divided into three storeys by partitions of green mortar; the lower storey was occupied by a female, the two others by males, with smaller cocoons.

I came to an even more remarkable example. Two Anthidia of my district, A. septemdentatum, LATR., and A. bellicosum, LEP., adopt as the home of their offspring the empty shells of different snails: Helix aspersa, H. algira, H. nemoralis, H. caespitum. The first-named, the Common Snail, is the most often used, under the stone-heaps and in the crevices of old walls. Both Anthidia colonize only the second whorl of the spiral. The central part is too small and remains unoccupied. Even so with the front whorl, the largest, which is left completely empty, so much so that, on looking through the opening, it is impossible to tell whether the shell does or does not contain the Bee's nest. We have to break this last whorl if we would perceive the curious nest tucked away in the spiral.

We then find first a transversal partition, formed of tiny bits of gravel cemented by a putty made from resin, which is collected in fresh drops from the oxycedrus and the Aleppo pine. Beyond this is a stout barricade made up of rubbish of all kinds: bits of gravel, scraps of earth, juniper-needles, the catkins of the conifers, small shells, dried excretions of Snails. Next come a partition of pure resin, a large cocoon in a roomy chamber, a second partition of pure resin and, lastly, a smaller cocoon in a narrow chamber. The inequality of the two cells is the necessary consequence of the shape of the shell, whose inner space gains rapidly in width as the spiral gets nearer to the orifice. Thus, by the mere general arrangement of the home and without any work on the Bee's part beyond some slender partitions, a large room is marked out in front and a much smaller room at the back.

By a very remarkable exception, which I have mentioned casually elsewhere, the males of the genus Anthidium are generally larger than the females; and this is the case with the two species in particular that divide the Snail's spiral with resin partitions. I collected some dozens of nests of both species. In at least half the cases, the two sexes were present together; the female, the smaller, occupied the front cell and the male, the bigger, the back cell. Other cells, which were smaller or too much obstructed at the back by the dried-up remains of the Mollusc, contained only one cell, occupied at one time by a female and at another by a male. A few, lastly, had both cells inhabited now by two males and now by two females. The most frequent arrangement was the simultaneous presence of both sexes, with the female in front and the male behind. The Anthidia who make resin-dough and live in Snail-shells can therefore alternate the sexes regularly to meet the exigencies of the spiral dwelling-house.

One more thing and I have done. My apparatus of reeds, fixed against the walls of the garden, supplied me with a remarkable nest of the Horned Osmia. The nest is established in a bit of reed 11 millimetres wide inside. (.429 inch—Translator's Note.) It comprises thirteen cells and occupies only half the cylinder, although the orifice is plugged with the usual stopper. The laying therefore seems here to be complete.

Well, this laying is arranged in a most singular fashion. There is first, at a suitable distance from the bottom or the node of the reed, a transversal partition, perpendicular to the axis of the tube. This marks off a cell of unusual size, in which a female is lodged. After that, in view of the excessive width of the tunnel, which is too great for a series in single file, the Osmia appears to alter her mind. She therefore builds a partition perpendicular to the transversal partition which she has just constructed and thus divides the second storey into two rooms, a larger room, in which she lodges a female, and a smaller, in which she lodges a male. She next builds a second transversal partition and a second longitudinal partition perpendicular to it. These once more give two unequal chambers, stocked likewise, the large one with a female, the smaller one with a male.

From this third storey onwards, the Osmia abandons geometrical accuracy; the architect seems to be a little out in her reckoning. The transversal partitions become more and more slanting and the work grows irregular, but always with a sprinkling of large chambers for the females and small chambers for the males. Three females and two males are housed in this way, the sexes alternating.

By the time that the base of the eleventh cell is reached, the transversal partition is once more almost perpendicular to the axis. Here what happened at the bottom is repeated. There is no longitudinal partition; and the spacious cell, covering the whole diameter of the cylinder, receives a female. The edifice ends with two transversal partitions and one longitudinal partition, which mark out, on the same level, chambers twelve and thirteen, both of which contain males.

There is nothing more curious than this mixing of the two sexes, when we know with what precision the Osmia separates them in a linear series, where the narrow width of the cylinder demands that the cells shall be set singly, one above the other. Here, the Bee is making use of a tube whose diameter is not suited to her work; she is constructing a complex and difficult edifice, which perhaps would not possess the necessary solidity if the ceilings were too broad. The Osmia therefore supports these ceilings with longitudinal partitions; and the unequal chambers resulting from the introduction of these partitions receive females at one time and males at another, according to their capacity.



CHAPTER 5. PERMUTATIONS OF SEX.

The sex of the egg is optional. The choice rests with the mother, who is guided by considerations of space and, according to the accommodation at her disposal, which is frequently fortuitous and incapable of modification, places a female in this cell and a male in that, so that both may have a dwelling of a size suited to their unequal development. This is the unimpeachable evidence of the numerous and varied facts which I have set forth. People unfamiliar with insect anatomy—the public for whom I write—would probably give the following explanation of this marvellous prerogative of the Bee: the mother has at her disposal a certain number of eggs, some of which are irrevocably female and the others irrevocably male: she is able to pick out of either group the one which she wants at the actual moment; and her choice is decided by the holding capacity of the cell that has to be stocked. Everything would then be limited to a judicious selection from the heap of eggs.

Should this idea occur to him, the reader must hasten to reject it. Nothing could be more false, as the merest reference to anatomy will show. The female reproductive apparatus of the Hymenoptera consists generally of six ovarian tubes, something like glove-fingers, divided into bunches of three and ending in a common canal, the oviduct, which carries the eggs outside. Each of these glove-fingers is fairly wide at the base, but tapers sharply towards the tip, which is closed. It contains, arranged in a row, one after the other, like beads on a string, a certain number of eggs, five or six for instance, of which the lower ones are more or less developed, the middle ones half-way towards maturity, and the upper ones very rudimentary. Every stage of evolution is here represented, distributed regularly from bottom to top, from the verge of maturity to the vague outlines of the embryo. The sheath clasps its string of ovules so closely that any inversion of the order is impossible. Besides, an inversion would result in a gross absurdity: the replacing of a riper egg by another in an earlier stage of development.

Therefore, in each ovarian tube, in each glove-finger, the emergence of the eggs occurs according to the order governing their arrangement in the common sheath; and any other sequence is absolutely impossible. Moreover, at the nesting period, the six ovarian sheaths, one by one and each in its turn, have at their base an egg which in a very short time swells enormously. Some hours or even a day before the laying, that egg by itself represents or even exceeds in bulk the whole of the ovigenous apparatus. This is the egg which is on the point of being laid. It is about to descend into the oviduct, in its proper order, at its proper time; and the mother has no power to make another take its place. It is this egg, necessarily this egg and no other, that will presently be laid upon the provisions, whether these be a mess of honey or a live prey; it alone is ripe, it alone is at the entrance to the oviduct; none of the others, since they are farther back in the row and not at the right stage of development, can be substituted at this crisis. Its birth is inevitable.

What will it yield, a male or a female? No lodging has been prepared, no food collected for it; and yet both food and lodging have to be in keeping with the sex that will proceed from it. And here is a much more puzzling condition: the sex of that egg, whose advent is predestined, has to correspond with the space which the mother happens to have found for a cell. There is therefore no room for hesitation, strange though the statement may appear: the egg, as it descends from its ovarian tube, has no determined sex. It is perhaps during the few hours of its rapid development at the base of its ovarian sheath, it is perhaps on its passage through the oviduct that it receives, at the mother's pleasure, the final impress that will produce, to match the cradle which it has to fill, either a female or a male.

Thereupon the following question presents itself. Let us admit that, when the normal conditions remain, a laying would have yielded m females and n males. Then, if my conclusions are correct, it must be in the mother's power, when the conditions are different, to take from the m group and increase the n group to the same extent; it must be possible for her laying to be represented as m-1, m-2, m-3, etc. females and by n+1, n+2, n+3, etc. males, the sum of m+n remaining constant, but one of the sexes being partly permuted into the other. The ultimate conclusion even cannot be disregarded: we must admit a set of eggs represented by m-m, or zero, females and of n+m males, one of the sexes being completely replaced by the other. Conversely, it must be possible for the feminine series to be augmented from the masculine series to the extent of absorbing it entirely. It was to solve this question and some others connected with it that I undertook, for the second time, to rear the Three-horned Osmia in my study.

The problem on this occasion is a more delicate one; but I am also better-equipped. My apparatus consists of two small, closed packing-cases, with the front side of each pierced with forty holes, in which I can insert my glass tubes and keep them in a horizontal position. I thus obtain for the Bees the darkness and mystery which suit their work and for myself the power of withdrawing from my hive, at any time, any tube that I wish, with the Osmia inside, so as to carry it to the light and follow, if need be with the aid of the lens, the operations of the busy worker. My investigations, however frequent and minute, in no way hinder the peaceable Bee, who remains absorbed in her maternal duties.