Contact Chemoreception, Mating Behaviour, and Sexual Isolation in the Rotifer Genus Brachionus

Although rotifers belonging to the order (or suborder) Ploima are predominantly parthenogenetic, sexuality often enters into their life histories. The transition to sexuality is brought about by rather ill-defined environmental conditions which induce amictic females to produce mictic female offspring. Both amictic and mictic females are diploid, similar in appearance, and lay eggs which develop partheno-genetically, but whereas amictic eggs are large, diploid, and always develop into females, mictic eggs are much smaller, haploid, and develop into males (Fig. 1). If males inseminate young mictic females which have not yet ovulated, mictic eggs may become fertilized and develop into large, thick-walled, diploid resting eggs (Fig. 1). These eggs, which invariably develop into amictic females, may immediately undergo further development and hatch or may remain dormant for many months before resuming development.

Resting eggs are resistant to desiccation and temperature extremes, and thus their adaptive significance is unquestionable. Ruttner-Kolisko (1949), in fact, believed that resting egg production per se is the only biological function of sexuality in ploimate rotifers. Although she did not think that genetic recombination is important, the fact that the genetically ideal system of reproduction is one involving an alternation of asexual and sexual organisms (Lewontin, 1957) suggests that sexuality in this group of animals has at least two potential biological functions. The possibility, however, that mictic female production in some ploimate rotifers may be a vestigial and hence non-functional characteristic cannot be excluded. Indeed, certain rotifer species have only rarely, if ever, been observed to exhibit periods of bisexuality.

If sexuality is to be a functional part of the life history of a particular rotifer species, the males and females of this species must be able to mate with each other. When considering the great number of ploimate rotifers which live together in the limnetic zone of ponds and lakes, where there are few, if any, habitat differences which might tend to separate the many species which are found together, the most efficient way for them to achieve a high degree of gamete conservation would appear to be through relatively specific mating processes. It is possible, though, that rather unspecific matings might be sufficiently successful in view of the fact that periods of bisexuality in many species are associated with extremely high densities of both males and females of the same species.

There is practically nothing in the literature which gives any clue to an under-standing of the mating process in ploimate rotifers. Wesenberg-Lund (1939, p. 225) recorded some casual observations on the mating behaviour of ploimate rotifers in general, but he gave neither a satisfactory description nor any indication of the specificity of the mating process in any one particular species. Whitney (1913) noticed that males of an unidentified species of Asplanchna would copulate with both males and females of this species but made no attempt to determine whether these males would mate with individuals of other species. The present study apparently represents the first attempt at a comprehensive study of the mating process of a ploimate rotifer.

The group of rotifers with which this paper is concerned is the very common ploimate genus Brachionus. The males of this genus, as of other ploimate genera, are considerably smaller than the females and are greatly reduced structurally (Fig. 1). They are adapted primarily for reproduction, and hence also for locomotion, and lack all those structures associated with food intake and digestion. They do not grow after hatching and have a life span roughly one-quarter that of females.

The three species which were available for this study were all collected in small, slightly alkaline ponds near New Haven, Connecticut. Two of these, B. calyciflorus tallas and B. angularis Gosse, are planktonic forms ; the other species, B. quadridentatus Hermann, is sometimes planktonic but is more often found attached to submerged material. The mature females of B. calyciflorus and B. quadridentatus were approximately the same size (about 350μ), while those of B. angularis were very much smaller (about 130μ). The mictic and amictic females of this genus appear to be identical in external morphology, although Ahlstrom( 1940) believes there may be minor differences.

B. calyciflorus and B. angularis were cultured in the laboratory in a defined, inorganic medium (pH 7·3) and fed on Euglena gracilis, strain z, and Chlamydomonas reinhardi, respectively. Culture methods are reported in more detail in another paper (Gilbert, 1963). Syracuse, embryological, and U.S. Bureau of Plant Industry Model watch glasses were used for maintaining cultures containing 10−20, 3−6, and 0·75−1·5 ml. medium, respectively.

Mictic female and consequent male production was induced at will in cultures of both B. calyciflorus and B. angularis by crowding amictic females (Gilbert, 1963). This was simply done by inoculating fresh culture medium with about 5·10 amictic females per ml. No B. quadridentatus males were available for this study because of the inability of this species to survive for more than a few days under the culture conditions used for either of the other two species.

The rotifers used which were not cultured in the laboratory were studied shortly after they were collected from the field ; these were the females of B. quadridentatus, Synchaeta sp., and Euchlanis sp.

A dissecting microscope was used with substage illumination to make all observations. Almost all manipulations involving rotifers were carried out under the microscope. Rotifers were transferred from one vessel to another with micropipettes.

For the mating behaviour of B. calyciflorus and B. angularis to be fully appreciated, it is worth while first to mention something about the general behaviour of these species. Although the females sometimes attach themselves by their feet either to the bottom or sides of the vessel in which they are living or to the surface of the culture medium, they usually swim about at a relatively steady speed. The direction of their movements appears to be random except in the presence of light gradients, in which case, being clearly positively phototactic, they quickly swim towards the area of greater light intensity and remain there. The males, which have no functional foot, and hence cannot attach themselves, swim about continually, usually at a considerably greater speed and in a straighter path than the females. The orientation of males towards light is not immediately obvious but is apparently both less precise than, and somewhat different in nature to, that exhibited by the females. At any rate, under the lighting conditions used to study these animals, both males and females moved at random with respect to their physical environment. Excluding occurrences of actual contact, they also moved at random with respect to one another, showing no tendency to be either attracted or repelled by members of either sex.

Matings between males and females of the same species were very frequently observed in both B. calyciflorus and B. angularis. Within either of these species no variations in this process were noticed. Males and females observed both very shortly after collection from several different ponds and also after several months’ culture in the laboratory all behaved in exactly the same manner. There was essentially no difference in the mating behaviour of these two species except that the reactions were generally more vigorous in B. calyciflorus than in B. angularis.

The female plays a completely passive role in the mating process. It is the male who has the responsibility of’finding’ a mate. He swims about in a normal, random fashion until he happens to come into actual head-on contact with the body of a female ; then he immediately exhibits a characteristic mating reaction. He arches his dorsal surface convexly and moves at variable speeds round and round the female, keeping one and often both of the extremities of his body in contact with her (Fig. 2), until he either loses contact and swims away or copulates. If the male does not contact the female head-on, no mating reaction ensues. Before a male loses contact with a female, the precopulatory mating behaviour may last from a fraction of a second to almost a minute, the time ordinarily depending mostly on the age and vigour of the male but also on the age of the female. Males which are quite old, e.g. older than about 8 hr. at 25°C., often do not react at all; the younger the male the more likely he is to exhibit a strong reaction and the greater the chance that this reaction will lead to copulation. Similarly, young females generally seem to induce male reactions which are both more intense and more likely to culminate in copulation than those elicited by very old females. Males never initiate mating reactions when they bump into other males, and females do not react to contact with either males or other females.

During copulation the male pierces the body wall of the female with his penis and injects the sperm into the body cavity. The site of injection, especially in older females, is usually the corona, which offers the largest most easily penetrable area, since it is not covered by the lorica. In young females, which have a softer lorica, the site of injection may be more variable. After the male has introduced the sperm, he attempts to pull away from the female. Copulating males are so firmly attached to the females that they often drag the females behind them for some time in their efforts to break loose and swim away.

To obtain some information on the specificity of the mating reaction, a series of tests using a combination of two procedures was carried out. In one procedure, males of one of the species of Brachionus were placed in a depression slide containing females of another species and were then observed when they came into contact with these females. If no reactions occurred after many contacts, the males were transferred into a depression slide containing females of their own species to test the possibility that they were impotent. In the other procedure, the males to be tested were placed in a depression slide containing females of both their own and another species and were then observed after they contacted a female. The results of these tests were clear-cut. Males of B. angularis exhibited definite mating reactions with both mictic and amictic females of their own species but displayed absolutely no indications of a reaction with females of B. calyciflorus and B. quadridentatus. Males of B. calyciflorus, in like manner, reacted with both mictic and amictic females of their own species but failed to react with females of Synchaeta sp., Euchlanis sp., Brachionus quadridentatus, and B. angularis. It is important to remark here, however, that a very small percentage of B. calyciflorus males gave very slight indications of starting to react with B. angularis females, but none of these males, which were in every case extremely active ones, ever actually began a definite mating reaction. The possibility that the relatively small size of the B. angularis females might have prevented the B. calyciflorus males from reacting with them was eliminated by the observation that these males will mate perfectly well with newly hatched females of their own species, which are considerably smaller than mature B. angularis females. It may be concluded, then, that the mating process in both B. angularis and B. calyciflorus is completely species-specific and that this specificity depends on the fact that only when the males come into head-on contact with females of their own species do they perceive a stimulus which causes them to initiate a mating reaction.

To gain some insight into the nature of the species-specific stimulus present in these female rotifers, males of B. calyciflorus were tested both with dead and with heat-killed females and also with fragments of females of the same species. In each case, the males exhibited characteristic contact reactions. These males, however, underwent no such reactions with dead or heat-killed females of B. angularis.

To test the likely possibility that the stimulus might be a chemical one, glass beads (about i mm. or less in diameter) were immersed in a drop of culture medium containing about 10 finely ground B. calyciflorus females. The suspension was allowed to evaporate on the beads, which were then placed in depression slides with medium containing B. calyciflorus males. The males exhibited typical mating reactions when they swam into these beads. Uncoated control beads had no effect on the males. B. calyciflorus males did not react with beads coated with the residue of similar suspensions made from either B. angularis females or B. calyciflorus males.

Since the stimulus which induces male mating reactions is evidently a chemical substance present in the bodies of females, the extraction of this substance appeared feasible. Both aqueous and ethyl ether extracts were prepared in the following manner. Mass cultures of B. calyciflorus were washed and highly concentrated by filtration through fine nylon mesh and then transferred to a depression slide. When, after the removal of enough water with a micropipette, the rotifers could be formed into a tacky mass, they were quickly and very thoroughly teased apart and ground with dissecting needles. Solvent (distilled water or anhydrous ethyl ether) was added to the depression, which was then scoured with dissecting needles to remove bits of rotifer fragments attached to the walls. The resulting suspension was stirred and centrifuged. All aqueous supernatants were, in addition, forced through 0·45μ Millipore filters. Small drops of extracts were evaporated on to glass beads, which were subsequently dried in a desiccator and placed in depression slides containing medium and also active B. calyciflorus males. These males intermittently exhibited typical mating reactions for as long as 15−30 min. on beads coated with residues from aqueous extracts but failed to react at all with beads coated with residues from ether extracts.

To find out if this apparently water-soluble chemical stimulus was permanently bound to female protoplasm or was released in some way into the environment, both B. calyciflorus and B. angularis males were transferred from fresh medium into filtrates of medium which had been highly conditioned by females of their own species. On transfer, these males immediately exhibited an unmistakable behavioural reaction. They swam about in a very small area, roughly one-ninth of that covered by normally swimming males in an equal period of time, continually describing small circles and loops and often stopping to reverse their direction (Fig. 3). Although this was the typical form of the reaction, it should be noted that occasionally males reacted by suddenly becoming almost motionless. Either form of the reaction decreased in intensity after 15−30 sec. and was no longer perceptible after about 2 min. This reaction was, with a few minor exceptions, extremely similar in males of both B. calyciflorus and B. angularis. B. angularis males swam abnormally slowly shortly after transfer from mass cultures into fresh media and then resumed normal speeds when transferred back to conditioned media. B. calyciflorus males did not exhibit this orthokinetic response. Furthermore, as in the mating reaction with females, B. calyciflorus males reacted more vigorously than B. angularis males.

It is quite clear that the reaction of males in female-conditioned media is a real mating reaction for the following reasons :

• Males exhibited the same type of reaction when transferred from fresh medium into aqueous female extract as they did when placed from fresh medium into female-conditioned medium.

• Beads coated with the residue of conditioned medium induced definite male contact reactions but lost their activity extremely rapidly, as the coating dissolved.

• In experiments in which media were conditioned by both B. calyciflorus and B. angularis, such that they were active enough to induce reactions in males of their own species, and were then tested with the heterologous species, no reactions were observed. This indicates that the active substance present in conditioned media possesses the same species-specificity as the active substance in both aqueous extracts of females and in females themselves. It should be mentioned, however, that an extremely active B. angularis-conditioned medium was made which occasionally induced almost normal reactions in B. calyciflorus males. This is perhaps in agreement with the previously mentioned observation that B. calyciflorus males would very occasionally exhibit incipient reactions with B. angularis females.

• When placed in female-conditioned media, the dorsal surface of the male became arched convexly, just as in the mating reaction on females. This is undoubtedly the reason why the males swim around in circles. In fact, males which have been experiencing normal mating reactions often swim in exactly this manner for a short while just after they lose contact with the females.

• Reactions in female-conditioned media and with females are both induced when the males suddenly became exposed to a very marked increase in concentration of the active substance. This exposure normally occurs only when the males actually bump into females but can be produced artificially in the manner already described. The obvious explanation for the rapid decline in the response of males in conditioned media, as opposed to the continued ability of males to react with females, is that the males experience a sudden increase in concentration of the substance only immediately after they have been transferred to the conditioned media. Thereafter they become adapted to the new concentration and resume normal behaviour unless exposed to another sudden increase in concentration, such as when, even in very highly conditioned media, they come into contact with an actual female. Males failed to react when exposed to gradually increasing concentrations of the active substance. Females did not show any change in behaviour when transferred from fresh to conditioned media, and males did not exhibit any reaction when transferred from conditioned to fresh media.

It is clear that the stimulus responsible for the induction of male mating behaviour is present only in the females of the same species, is species-specific, and is chemical in nature. The chemical substance can be extracted from females with water, and there is considerable evidence that this same substance is released into the environment by growing females. The possibility of obtaining some chemical characterization of this substance, by treating aqueous extracts or female-conditioned media in various ways, was now open.

The results of the tests performed on both extracts and on conditioned media of B. calyciflorus are summarized in Table 1. Details, both of the bioassays and treatments used, and of the results obtained, are presented below.

Active aqueous extracts were prepared in the manner already described by using about 300 crushed rotifers per ml. of extract. These could be stored in glass vials at 5°C. for several days without significant loss of activity. If the extract could be treated in such a way that it would not be rendered toxic to males, it was so treated and then coated on beads for bioassay. If the treatment necessitated using a toxic chemical, previously coated beads were immersed in this toxic solution, washed with distilled water, and then tested on males. The bioassay by the mating reaction with coated beads could not be strictly quantified, but the intensities of the reactions with coated experimental and control beads could be compared. Control and experimental beads were always tested in separate depressions, and the same males were used for both.

Extracts were placed in glass vials which were immersed in boiling water. Samples were withdrawn after 1, 5, 10, 15, 30, 45 and 60 min., coated on Beads, and assayed. Activity gradually diminished so that it was no longer perceptible after 30−45 min.

Extracts were treated and bioassayed in the same way as the female-conditioned media and found to be equally resistant. See section on conditioned media for details.

Extract (0·5 ml.) was dialysed against distilled water (2000 ml.) for 26 hr. at 5°C. The dialysed extract was coated on beads and found to be as active as control extracts kept under similar conditions without dialysis.

Extract was prepared in 0·006 M-NaH₂PO₄-Na₂HPO₄ buffer (pH 7-7). A 0·5 % solution of trypsin (lyophilized, General Biochemicals Inc.) was also prepared in this buffer. The following mixtures were then made and incubated for 60 min. at 36°C.: (1) 9 parts extract: 1 part trypsin solution, (2) 9 parts extract: 1 part buffer, (3) 9 parts buffer: 1 part trypsin solution, and (4) 9 parts heat-deactivated extract : 1 part trypsin solution. After incubation, each mixture was evaporated on beads and assayed. Mixtures (1) and (2) were equally active; mixtures (3) and (4) induced no male reaction.

Coated beads were immersed in 0·1N-HC1O₄ for 30 min. and in 0·6 N-HC10₄ for 5 min. at 25°C., washed, and assayed. They were just as active as control beads, which were treated similarly but with distilled water.

Active conditioned media were obtained by crowding young females in fresh medium (5−10 animals per ml.), feeding them an excess of Euglena (10−15 thousand cells per ml.), and letting them grow for 24 hr. at 25° C. The conditioned media were centrifuged, and the supernatants were forced through either 0·45 or 0·8oμ Millipore filters. Like the aqueous extracts, conditioned media could be stored in glass vials for several days at 5°C. without appreciable loss of activity. The bioassay for the active substance in conditioned media, like the bioassay in aqueous extracts, could be quantified only by comparing the intensities of the male reactions in treated and control preparations. The assay in conditioned media, however, was much preferable to that on coated beads because there was no danger of false negatives due to the substance merely dissolving off the beads.

The assays in conditioned media were performed in depression slides by pipetting males from fresh media into those to be tested. Several different males were used for each assay, and the same males were used to test both treated and control media. Although the results of most assays were very easy to score, sometimes considerable care and time had to be taken. Some treated media were toxic to the males and induced a reaction in which the males would spin around very rapidly and which might be misinterpreted as a mating reaction. This response to toxic conditions could always be distinguished from the mating reaction by putting males already adapted to conditioned medium into the treated medium ; if a reaction occurred, it was not a mating reaction. Similarly, if a very slight mating reaction was suspected, males could be transferred from the medium in question to one known to possess definite activity; if no reaction was observed, there was good reason to believe that the questionable reaction was a real but weak mating reaction. Since males used in these assays were observed just after they left the mouth of the micropipette, it was standard procedure to pipette them in and out of fresh medium several times just before transferring them to the medium to be tested. In this way, the behaviour of the males in both fresh and treated media could be most accurately compared.

Active media were placed in glass vials which were immersed in boiling water. No significant loss of activity was noticed after 1 hr. ; after 3 hr. a definite loss of activity had occurred.

The following two mixtures were prepared: (1) 8 parts active medium: 2 parts 0·1 N-NaIO₄ and (2) 8 parts active medium:2 parts distilled water. Both mixtures had similar pH values (about 5-5). Each was incubated at 29°C. for 7 hr. in the dark, and assayed directly ; they were found to possess equal activity. In another experiment an active medium in 0·01 N-NaIO₄ and an equivalently diluted active control medium were incubated at 27°C. for 23 hr. in the dark and then assayed. Both mixtures contained definite and equal activity.

Active media (0·2 ml.) were dialysed against fresh media (400 ml.) for 25 hr. at 9·5°C. The dialysed media were inactive. Media kept under similar conditions without dialysis retained activity. Fresh medium (4 ml.) was placed in dialysis tubing and immersed in active conditioned medium (15 ml.). After 27 hr. at 25°C. the medium within the tubing was assayed for the first time and found to possess definite activity.

The residue from 1 ml. of active medium was suspended in 3 ml. of anhydrous ethyl ether for several minutes at 25°C. The suspension was centrifuged, and the supernatant was filtered through glass wool and evaporated to dryness. The residue was dissolved in 1 ml. of fresh medium. The resulting solution possessed no activity. When the ether-insoluble residue was redissolved in 1 ml. of distilled water, the resulting solution was active.

One ml. of active medium was thoroughly mixed with 4 ml. of ethyl ether for several minutes at 25°C. When the aqueous phase was separated from the ether phase and gently heated to drive off the remaining ether, it was assayed and found to be just as active as untreated control medium.

Active media adjusted to 0·01 N-HCI (9 parts medium: 1 part 0·1 N-HC1) were kept at 25°C. for 2 hr., neutralized with 0·1 N-NaOH, and then assayed. No loss of activity was ever detected. It should be noted that extreme care had to be taken when neutralizing media; if the pH was slightly too acidic or basic, males underwent a reaction indicative of toxic conditions.

Active media adjusted to 0·01, 0·1, and 1·2 N-HCI were incubated at 100°C. for 60 min. along with active control media equivalently diluted with water. After heating, the medium in o-oi N-HCI was simply neutralized and assayed. Since concentrations of NaCl much greater than 0·01 N were toxic to males, the media containing 0·1 and 1·2 N-HCI were evaporated to dryness to drive off most of the HC1. This was quickly done in depression slides set in a 60°C. water bath under jets of air. Control media were subjected to this same treatment. The residues were dissolved in distilled water (about one-half the volume of the original medium), neutralized with a small amount of 0·1 N-NaOH, and assayed. Media treated with 0·01 and 0·1 N-HCI retained definite activity, but activity was destroyed by 1 ·2 N-HCI. All control media were active.

Active medium was adjusted to 0·01 N-NaOH (9 parts medium : I part o·1 N-NaOH) and kept at 25°C. Samples were withdrawn after 1, 5, 15, 45 and 120 min., neutralized with o-i N-HCI, and assayed. A significant decrease in activity was noted after each time interval, such that there was only very slight activity after 45 min. and negligible activity after 120 min.

Active media were not destroyed by brief treatment with 0·7 N-NH₄OH. These media (9 parts medium: 1 part 7 N-NH₄OH) were left in depression slides at 25°C. for 60 min., after which time the pH had decreased to about 8-5. The experimental media were neutralized with 0·1 N-HCI and found to be just as active as the controls.

A 1 % solution of trypsin (lyophilized, General Biochemicals Inc.) was prepared in 0·04 M-NaH₂PO₄-Na₂HPO₄ buffer (pH 7·7). The following three mixtures were incubated at 36°C. for 2 hr. and then directly assayed : (1) 9 parts active medium : 1 part trypsin solution, (2) 9 parts active medium: 1 part buffer, and (3) 9 parts fresh medium: 1 part trypsin solution. Mixtures (1) and (2) possessed definite and equal activity; mixture (3) had no effect on males.

A 1 % solution of pepsin (3 × crystallized, General Biochemicals Inc.) was prepared in o-oi N-HCI. The active medium used in this experiment was adjusted to 0·01 N-HCI. The following four mixtures were incubated at 36°C. for 2 hr., neutralized with 0·1 N-NaOH, and then directly assayed: (1) 9 parts active medium:i part pepsin solution, (2) 8 parts active mediums parts pepsin solution, (3) 8 parts active medium:2 parts 0·01 N-HCI, and (4) 8 parts fresh medium in 0·01 N-HCI:2 parts pepsin solution. Mixtures (1), (2), and (3) possessed definite and equal activity; mixture (4) had no effect on males.

A 5 % suspension of carboxypeptidase (3 × crystallized, Nutritional Biochemicals Corp.) was dissolved in an equal volume of 0·1 N-NaOH and then immediately diluted with media, as the enzyme gradually loses activity at very high pH values (Rupley & Neurath, i960). The following three mixtures were incubated at 31°C. for 2 hr. and then directly assayed: (1) 23 parts active mediums parts 2-5 % carboxypeptidase in 0·05 N-NaOH, (2) 23 parts active medium : 2 parts 0·05 N-NaOH, and (3) 23 parts fresh medium:2 parts carboxypeptidase solution. The pH of these mixtures was about 7·8. Mixtures (1) and (2) possessed definite and equal activity; mixture (3) had no effect on males.

Active media (prepared normally, in 0·01 N-HCI, and in 0·01 N-NaOH) were mixed with very small amounts of Norit charcoal and Millipore-filtered; the filtrates were neutralized (when necessary), bioassayed, and found to possess no activity. Small amounts of Norit, previously washed with 50% ethanol and then water, were mixed with 1 ml. of both active and fresh medium. Both suspensions were passed through glass tubes plugged with a small amount of glass wool. One ml. of 50 % ethanol was then passed through each of these two tubes, collected in depression slides, and evaporated to dryness with heat. Both residues were dissolved in 0·5 ml. of fresh medium, and the solutions were assayed. Definite activity was eluted off the charcoal which had been previously mixed with the active medium; the control eluate had no effect on males.

Active media adjusted to 0·01 N-HCI were mixed in depression slides with the H+ form of the anionic resin Dowex 50 × 12 (1 part resin: i, 5, or 10 parts medium) for several minutes and then either Millipore-filtered or centrifuged. The filtrates or supernatants were neutralized and always found to be inactive. The treated resins were washed with distilled water and then mixed with 0-7 N-NH₄0H (two-thirds the volume of the original active medium). The mixtures were filtered into depression slides. After the eluates had been kept at 25°C. for about 45 min., they were neutralized, assayed, and found to possess definite activity. Active media adjusted to 0·01 N-HCI were treated with the Cl^- form of the cationic resin Dowex 1× 8 (1 part resin:1 or 4 parts medium), filtered, neutralized, and assayed. Activity remained in these filtrates.

Active media adjusted to 0·01 N-NaOH were mixed for several minutes in depression slides with the Na+ form of Dowex 50 (1 part resin:4 parts medium), Millipore-filtered, neutralized, and assayed. These filtrates were always just as active as control media kept in 0·01 N-NaOH for equal lengths of time before neutralization. Dowex i was partially converted to the OH^- form by equilibrating the Cl^- form of this resin in 0·01 N-NaOH. Active media adjusted to 0·01 N-NaOH were mixed with this resin (1 part resin: 14 parts medium), filtered, neutralized, and assayed. The filtrates contained significantly less activity than the controls. The treated resins were washed with distilled water and then mixed with HC1 to bring the pH of the mixture down to about 2 (volume of acid water was about two-thirds that of the original active media). The mixtures were filtered, neutralized, and assayed. Definite activity was present in these eluates.

The three species of the genus Brachionus studied in this paper, B. calyciflorus, B. angularis, and B. quadridentatus, along with another species of this genus, B. urceolaris, which was unfortunately not found at the time of this study, all commonly occur together in alkaline ponds throughout the temperate regions of the world (Wesenberg-Lund, 1930; Buchner, 1941; Pejler, 1957; Gilbert, unpublished observations). All four of these species may be present in one body of water at the same time, but usually no more than two or three are found together. B. calyciflorus and B. angularis, the two most planktonic and closely associated species of this group, seem to occur together most commonly.

Although there is no apparent correlation between the appearance of sexual periods in the different species which are found together, they do quite often coincide or overlap one another. Wesenberg-Lund (1930) and I have both observed concurrences of sexual periods in B. calyciflorus and B. angularis, and Buchner (1941) has reported that these two species and also B. urceolaris may all exhibit bisexuality at the same time. The ability of these Brachionus species to exist together and also to attain simultaneously the high population densities apparently necessary for the induction of mictic female production (Gilbert, 1963) is probably due to their ability to avoid mutual competition. B. angularis and B. calyciflorus ingest organisms of different size ranges, while B. urceolaris and B. quadridentatus, being less planktonic, can exploit areas usually unfrequented by the former two species. It is thus not difficult to envisage naturally occurring situations where large numbers of both males and females from several species of this genus are all swimming about together.

The present study has demonstrated that B. calyciflorus and B. angularis are sexually isolated from each other and from all other ploimate species tested (B. quadridentatus, Synchaeta sp., and Euchlanis sp.). Interspecific mating attempts are never observed. This complete isolation is based on the fact that the males of these two species only initiate mating behaviour when chemoreceptors, located somewhere on the corona, come into contact with, or are exposed to, a sudden increase in concentration of a species-specific substance present only in the females of their respective species. Hence, males never attempt to mate with other males of their own species. Upon perception of this releaser substance, a reflex, in which the dorsal surface of the male becomes arched convexly, is immediately triggered. The arching of the back is probably the manifestation of an attempt to pierce the body wall of a female with the penis. As a consequence of this reflex, the male moves round and round the female until he either loses contact and swims away or impregnates the female. The female does not take an active part in the mating process. Male mating behaviour is completely innate and is not reinforced by learning. Unmated males react just as well as, if not better than, experienced males and cannot be made to mate with females of species other than their own, no matter how often they are exposed to them.

No attempt was made to determine the location of the releaser substance within the female. Males seem able to perceive it over the entire body surface, as they appear to react indiscriminately on contact. Whether the substance is localized in the lorica itself, present only in the soft body parts and perceived through the lorica, or secreted on to the outside of the lorica is not known. Either of these arrangements could explain the observation that young females induce more intense mating reactions than old females. As the female grows the substance could become less easily perceived; it might be mechanically masked as the lorica becomes thicker and harder with age, or it might be secreted in smaller amounts as the female matures. At any rate, the fact that young females are more likely to be inseminated than old ones is highly adaptive, because only in these females can mictic eggs be fertilized and develop into resting eggs.

The releaser substance is extractable from females with water. The activity in aqueous extracts is non-dialysable, heat-stable, not extractable with ether, and resistant to trypsin, periodate oxidation, and dilute perchloric acid. Females release into the medium, probably by secretion or excretion, a substance which gives every indication of being identical with the substance present in the aqueous extracts. It also elicits species-specific mating reactions and is heat-stable, water-soluble, insoluble in ether, and resistant to trypsin and periodate oxidation. It is, however, dialysable. These observations may be interpreted on the assumption that living females release a small-molecular active substance which is stored within their bodies in some bound, non-dialysable form. In addition to the chemical properties already mentioned, the dialysable form of the active substance is quite acid-stable, withstanding treatment of 0·1 N-HCI for 60 min. at 100°C. It is base-labile, resistant to pepsin and carboxypeptidase, and adsorbed on charcoal. Its retention and release by ion-exchange resins indicates a positive charge in very acid solution and a negative charge in very basic solution.

On the basis of these properties, little can be said about the possible chemical nature of the releaser substance. It is certainly not a protein, but it might be a peptide having an arrangement of amino acids resistant to the peptidases used. The ability of the substance to withstand prolonged periodate oxidation and to become positively charged in acid solution suggests that it is probably not a polysaccharide. The substance does not possess lipid properties and cannot be a simple fatty acid, which would exhibit some solubility in ether and would be stable to hydrolysis in 1 ·2 N-HCI. It is also most likely not a simple nucleotide, for these would be strongly adsorbed on Dowex 1 and stable in alkali. In summary, all that can be concluded about this substance is that it is a small molecule, probably with a molecular weight less than 4000, which appears to be amphoteric and which, as indicated by its adsorption on charcoal, probably possesses an aromatic or heterocyclic ring.

To gain some additional information about the chemical nature of biological specificity, it would be of great interest to determine the molecular structures of the speciesspecific releaser substances present in the females of different Brachionus species. It is doubtful, however, whether much further progress can be made using unfractioned female-conditioned media as a test material. Chromatographic and electrophoretic studies using highly concentrated conditioned media might be practicable and very informative, but the isolation of the releaser substance in pure form would be most desirable. The fact that the substance is dialysable and can be easily adsorbed on, and eluted off, both charcoal and Dowex 50 would be useful for this purpose.

Other examples of animals which depend, at least to some extent, on contact chemoreception for recognizing males are quite uncommon in the literature. Kaston (1936) has shown that the males of lycosid spiders are aided in recognizing appropriate mates by sensing the presence of an ether-soluble substance on the integument of the female. There is no evidence about the specificity of this substance. Similarly, Roth & Willis (1952) have shown that males of the cockroach Blatella germánica initiate their courting behaviour when certain chemoreceptors come into contact with a chloroformsoluble substance on the bodies of females and young males. This mechanism of species recognition, however, is not absolutely specific, for males will sometimes court females belonging to other genera. Some sort of contact recognition seems to be operative in the copepod Cyclops and in the isopod Asellus. Holmes (1909), for instance, has observed that males of C. fimbriatus will only temporarily clasp females of other species, and Vandel (1926) has reported that males of several species of Asellus may grasp females of species other than their own for as long as several days but generally reject them before inseminating them. It would be interesting to investigate these cases in greater detail. The highly specific and immediately effective mechanism of species and mate recognition demonstrated in the genus Brachionus is apparently without parallel. As far as I am aware, comparable cases of recognition mechanisms based on contact chemoreception have not been reported in either freshwater invertebrates or invertebrates in general. Such cases, however, are probably quite common in nature.

It is to be expected, for example, that other rotifer genera will have similar ways of insuring the specificity of their matings. A mechanism, whereby such specificity might be attained through the mechanical inability of a particular species to consummate mating attempts with a species other than its own, would be very unlikely. The basis of this kind of species isolation is generally the existence of interspecific differences in copulatory apparatus. The very simple and rather crude nature of the copulatory act in rotifers—hypodermic impregnation—would suggest that males could easily impregnate females of many different genera. Therefore some sort of sexual isolating mechanism would appear to be the best way that co-occurring ploimate rotifers could reproductively isolate themselves from one another.

Since the female rotifer probably never takes an active part in the mating process, the method of sexual recognition by which, for example, several water mites of the genus Piona achieve some degree of sexual isolation would be impossible. In these animals the female rejects males of species other than their own by actively disentangling herself from the grasp of the male (Mitchell, 1957). Parker (1901) observed that males of the copepod Labidocera aestiva seem to be chemotactically attracted toward the females of this species, but it is unlikely that such a mechanism could facilitate proper pair formation in rotifers. In the first place, it is extremely improbable that a gradient of a sex attractant from a planktonic organism could ever naturally exist in a turbulent body of water. In the second place, it would seem impossible for an animal incapable of image formation, such as a rotifer, to be able to distinguish the individuals emitting the sex attractant from the many other individuals which are swimming in exactly the same area and which are emitting no attractant. Thus, it would not be surprising to find that other rotifers besides Brachionus have evolved a method of species and mate recognition based on contact chemoreception, since such a method appears to be the only one within the structural capabilities of these relatively simply constructed animals that might be effective. It is interesting in this respect to note that many ciliates, faced with more extreme structural limitations, have evolved a somewhat analogous and equally specific mechanism for mate recognition. In Paramecium bursaria this mechanism has recently been shown to depend on a surface reaction in which the cilia of complementary mating types adhere to each other (Siegel & Cohen, 1962). The mating-type substance within the cilia appears to be a protein.

Although it seems probable that all rotifers exhibiting sexual periods will have some mechanism which serves to confine their sexual activity and that this mechanism will involve contact chemoreception, only further studies will show to what extent these predictions are valid. The observations of Whitney (1913), indicating that males of a species of Asplanchna will mate with other males of this species, provide some evidence that Asplanchna and Brachionus possess at least somewhat different mechanisms of mate recognition. The mating process in the genus Asplanchna, however, must be studied much more thoroughly before any comparisons between these two genera can be made.

This paper is part of a dissertation submitted in partial fulfilment of the requirements for the degree of Doctor of Philosophy at Yale University, New Haven, Connecticut. I have greatly appreciated the advice and encouragement of Prof. G. Evelyn Hutchinson and am indebted to Dr Gerard R. Wyatt and Dr George Brawer-man, without whose counsel much of the chemical portion of this study would have been impossible. I wish to thank Dr John L. Brooks and my wife, Caroline, for reading the manuscript.