Dugesia tigrina was reared under various combinations of long- or short-day length with three temperature levels, 20, 23 and 26 °C. The resulting asexual and sexual reproduction was recorded.

Animals reared under short days exhibited shorter generation length and a higher number of worms hatching per cocoon than animals reared under long days.

Changes from short to long day or vice versa, experienced by animals during development, resulted in retardation or complete absence of sexual reproduction. Cocoons laid by these individuals showed a high degree of infertility.

D. tigrina was also collected in cocoons from the natural habitat shortly before hatching. When exposed to completely different light regimes they responded with a uniform generation length.

It is concluded that generation length, apart from temperature influences, is governed by photoperiod during early development of individuals, probably already in the maternal body. The number of worms hatching from one cocoon, likewise, is under photoperiod influence which acts at some time after hatching of the parent.

Our knowledge of photoperiod control over development and reproduction of the lower metazoa is both scanty and recent. Planarians, being among the most primitive Bilateria, have been widely investigated with respect to their regenerative abilities. They deserve more detailed investigation into their reproductive physiology, its regulation and control, especially since aspects of asexual reproduction and regeneration are so closely related.

A general review of the literature relating to the induction of sexual development in planarians was made by Vowinckel (1968, 1970b), and therefore is not repeated here.

It has recently been demonstrated (Vowinckel, 1970a, b) that changes in both photoperiod and temperature can independently induce sexual development in Dugesia tigrina. The effect of illumination appears to be profound and needs more detailed study. Our earlier work suggested that the inductive effect of photoperiod may take place early during development of a planarian.

We started with the assumption that exposure to short-day or long-day illumination combined with different levels of constant temperature should have different effects on reproduction. We report below on the results of experiments designed to test these assumptions.

Material. Two groups of approximately 75 sexually developed Dugesia tigrina were collected from a population in the St Lawrence River in the autumn of 1969 and maintained in the laboratory at a photoperiod of LD (light: dark) 16:8 at 20 °C. When cocoon laying started one population was transferred to LD 12:12 at 20 °C. The cocoons laid by these two populations provided the material for the populations of series 1 and 2 described below. They were collected 5 times per week and each series was fully established before the next was started. The age of individuals within each series varies by less than 2 weeks with the exception of the first long-day series at 20 °C (L120), which was the first series to be established and took 6 weeks to complete.

The cocoons that formed the five populations of series 3 were all collected in the natural habitat on the same day in July 1970.

Experimental design. Populations were established from cocoons. The latter were raised in six incubators which represented all possible combinations between long- and short-day illumination and three temperature levels. Long day was represented by LD 16:8 (On from 6 a.m. to 10 p.m. EST - eastern standard time). This approximates the daylength at the time of cocoon deposition and hatching of the wild population. LD 12:12 (On from 8 a.m. to 8 p.m. EST) was chosen to represent short days. It occurs in the natural habitat at a time when most worms have completed their sexual development. The lowest temperature level was 20 °C at which cocoon deposition takes place in the river, 23 °C typically represents the temperatures of the natural habitat during July and August and 26 °C approximates the highest temperatures which the natural population is likely to encounter.

There were three main series of populations: one series in which each population experienced only one type of daylength throughout the extent of the experiment, another series where populations were exposed to changes of day-length during their development and a third series of populations deriving from cocoons laid in the natural habitat.

Series 1 consisted of 11 populations (the twelfth could not be completed). Two parallel populations were raised under each of six temperature-daylength combinations in the following manner:

Since the cocoon donors for these experimental series were maintained at 20 °C the four populations at 20 °C experienced constant temperatures, while the seven populations at 23 and 26 °C experienced one initial change of temperature from 20 °C to 23 and 26 °C respectively on the day the cocoons were laid.

Series 2 consisted of 17 populations. These were populations which experienced a change in photoperiod régime during development in the cocoon or shortly after. Changes were from short to long day or vice versa and took place either (n) within 24 h after the cocoon was laid, or (b) within 24 h after the worms hatched from the cocoon, or (c) 2 weeks after hatching. Ideally, this series would have included populations for each of the six types of shifts (plus duplicates) for all three temperatures. However, the cocoon donors stopped laying and only the following populations were established:

All cocoons laid under experimental conditions by any population of series 1 and 2 were maintained separately under room temperature (air-conditioned, approximately 21 °C). Photoperiod was not controlled. The number of worms hatching from these cocoons were counted regularly and then discarded.

Series 3 consisted of five populations. These were all established from cocoons collected on the same day in the natural habitat. One population (L-Sf) was consequently maintained on a shaded float in the natural habitat until the age of 85 days (21 September 1970). It was then brought to the laboratory and installed in an incubator at 20 °C and LD 12:12. The other four populations were established directly in the laboratory. They were installed in incubators at 23 °C in the following manner:

The groups that were installed in incubators all experienced a retardation of On and Off times by approximately 1 h on the first day while the short-day population was exposed, at this time, to a shift in photoperiod from long day to short day. Worms began to hatch the day after arrival in the laboratory.

The photoperiod of the last two populations was reduced, after all worms had hatched, by twice weekly steps of 15 min, min in the morning and at night. When LD 12:12 was reached by the two L-Sn groups all five populations of series 3 were switched on the same day and in one step to 20 °C (21 September).

Maintenance. All populations were maintained in incubators at controlled temperatures and photoperiods.

Both temperature and light were continuously monitored by a multi-channel a.c.-d.c. recorder. Photoperiod was automatically regulated. The light source was a shaded 9 in (22·8 cm) fluorescent light of less than 750 lux intensity at 10 cm distance. Diurnal temperature variations stayed below ±0·5 °C. Due to the long span of the experiment (over 1 year) and consequent slow changes of room temperature, the overall variation was only just below ± 1·0°C. Since these changes affected all incubators in a parallel fashion we do not think that our results were much influenced by them.

Animals were maintained in glass containers at population densities not above 0·1 worm/ml in non-chlorinated river water (Vowinckel, 1970 a). All containers were covered with non-translucent material. Beef liver was fed to all populations 4 to 5 times weekly on an all or none basis. Cultures were cleaned on the same day and water replaced at the same temperature.

Collection of records. Population counts were taken approximately every 8 weeks. All additions to populations (fission products and cocoons) were recorded daily except at weekends and removed. All losses from the population were likewise recorded. They resulted mostly from individuals that crawled under the lid and dried up. Diseased animals were very rare. Animals that fissioned anterior to the pharynx were removed entirely and counted as losses.

Notes on the degree of sexual development of populations, as visible externally, were kept with the daily record and all individuals were submitted twice to an examination under the dissecting microscope during the first 80 days of their life. Cocoons from experimental series, after hatching, were investigated under the dissecting microscope for remaining individuals. Infertile cocoons were counted at the end of the hatching period.

Statistical evaluation. All records of fission products and cocoon deposition were recalculated and expressed as percentage of the presently existing population. They were then averaged (fission products) or summed (cocoons) over 5 days.

Since a population’s generation length, measured as time passed from laying of first parental to laying of first filial cocoon, is based on the performance of one or two individuals only the data for the two L20 populations were combined as were those of the two S20 populations, the difference not being significant. All cocoons laid up to 50% cocoon production (and therefore representing one cocoon laid for every second animal) were then used to evaluate the difference between L20 and S20 by a Mann-Whitney U-test after Siegel (1956).

A t-test was used to evaluate the significance of the difference between the number of worms hatched from all fertile cocoons laid under long-day conditions at 20 °C (including shifted series no. 2) as against those laid under short-day conditions at 20 °C.

Reproduction at 20 °C. The reproductive activity, both sexual and asexual, of all populations at 20 °C is summarized in Fig. 1 and Tables 1–3. All populations reproduced asexually, and all but one population reproduced sexually. The generation length of populations differs very obviously. When measured from the first parental to the first filial cocoon of each population, the two short-day (S20) groups have a generation length of 87 and 97 days. The value for the two long-day (L20) populations are 110 and 129 days. Since such a comparison is based on the performance of one or two individuals only, 50% cocoon production of the combined L populations was tested against the combined S populations by a Mann-Whitney U-test. Based on these statistics the difference in generation length between L20 and S20 populations is highly significant with a probability of less than 0·000001 that this difference arose accidentally. Since the differences in generation length between the S20 or L20 groups and the shifted populations are even greater, further U-tests were not calculated.

Table 1.

Cocoon production and generation length of groups at all temperatures

Cocoon production and generation length of groups at all temperatures
Cocoon production and generation length of groups at all temperatures
Fig. 1.

Asexual and sexual reproduction of groups at 20°C in percent of population producing one fission product or cocoon. Asexual: representing 5-day means, in white; sexual : representing 5-day sums, in black. Asterisk: end of record. The figures on the right indicate the size of each population. All L under long-day, all S under short-day illumination. Values below 1 % are entered as 1 %. Black bars are added above white bars.

Fig. 1.

Asexual and sexual reproduction of groups at 20°C in percent of population producing one fission product or cocoon. Asexual: representing 5-day means, in white; sexual : representing 5-day sums, in black. Asterisk: end of record. The figures on the right indicate the size of each population. All L under long-day, all S under short-day illumination. Values below 1 % are entered as 1 %. Black bars are added above white bars.

Approximately 50% of the cocoons laid by the series with constant photo-period were fertile with the exception of the second short-day population (S2) which for some unknown reason only laid sterile cocoons.

The generation length of populations subjected to a photoperiod change early in life differs considerably from that of populations kept at constant illumination. Groups shifted from long day to short day spent most or all of their post-hatching life under short-day conditions. However, their generation length in every case is much longer than that of the constant short-day populations. In each case cocoon deposition is postponed, by 8 and 4 weeks respectively, leading to generation lengths of 149 days (L-Sa, shifted at laying) and 124 days (L-Sb, shifted at hatching). When the shifting occurs two weeks after hatching (L-Sc) cocoon laying is postponed to 310 days, while the parallel population still had not laid after 325 days. Periodic evaluations of the two latter groups showed that nearly all individuals were fully developed sexually after 150 days.

A photoperiod shift in the other direction, from short day to long day, affected the generation time even more drastically. As compared with the constant long-day population cocoon deposition was retarded by approximately 18 weeks (generation length 253 days) when shifted at laying (S-La) and by approximately 10 weeks (generation length 191 days) when shifted at hatching (S-Lb). In addition, these cocoons were all sterile in contrast with those laid after an L-S shift (Table 1). In general, cocoon deposition was least retarded if the shifting took place at hatching time.

The populations of series 3 were all raised from cocoons laid in the natural habitat and under natural long-day conditions. They experienced completely different light régimes after hatching. In spite of this their generation lengths are remarkably uniform (Table 3).

One interesting side effect of the postponement of cocoon deposition after the reproductive organs have developed was the occurrence, in most worms, of supernumerary copulatory complexes. Up to four such complexes in the same animal have been observed several times, the hindmost being the smallest and the anteriormost the largest and functional. The production of a very small and empty cocoon from a supernumerary gonopore was observed. An impression of the generality of this effect can be gained by the following figures: Of 234 Dugesia tigrina (cocoons collected from the St Lawrence in July 1969) raised under LD 16:8 at 20 °C, 43·6 % had one gonopore, 44·9 % had two, 10·7 % had three and 0·9 % had four gonopores.

Asexual reproduction of cocoon-hatched populations is condensed in Table 2. A comparison shows that all populations at 20 °C that grew up under long-day conditions (including shifted groups) reached 50 % cocoon production much earlier than short-day populations with the exception of the first group at L20 which is the only population with a wide age range (see Material and Methods).

Table 2.

Asexual reproduction of all groups at 50 and 100% of population and on day 63, for all temperatures

Asexual reproduction of all groups at 50 and 100% of population and on day 63, for all temperatures
Asexual reproduction of all groups at 50 and 100% of population and on day 63, for all temperatures
Table 3.

Generation length of five populations all deriving from the same batch of cocoons collected in the natural habitat

Generation length of five populations all deriving from the same batch of cocoons collected in the natural habitat
Generation length of five populations all deriving from the same batch of cocoons collected in the natural habitat

Reproduction at 23 °C. After an initial few fission cycles worms in general at this temperature oscillated between sexual and asexual reproduction. A full set of reproductive organs including copulatory complex and filled seminal vesicles would develop and then be lost again by fission, this cycle to be repeated several to many times. Eventually, however, sexual reproduction asserted itself in most populations. This is summarized in Table 1. Reproduction was largely irregular, both in generation length and the number of cocoons laid, most of which were infertile. Only one population showed over 100 % cocoon production with 12 % of these fertile.

As was to be expected asexual reproduction of cocoon-hatched groups was considerably higher at 23 °C than at 20 °C. It also started earlier (Table 1). At this temperature there was no difference in fission rate between long- and short-day series.

Reproduction at 26 °C. This temperature corresponds to the highest summer values encountered in the natural habitat from which our stock is derived. No external sexual development was observed in any population under these conditions. After very high initial fission rates (Table 2) for approximately 2 months the constant long-day group (L, no shifts) showed signs of exhaustion. The fission rate declined, some worms developed very long tails (tail twice as long as body anterior to fission zone) and many began to fission anterior to the pharynx, often at ovary level. After a month of this abnormal behavior worms began to die without evident signs of disease. At this point all 26 °C populations, including those exposed to photoperiod changes were shifted to 20 °C to see if they were still capable of sexual reproduction. Their record (Table 1) shows that all populations laid some cocoons but that with the exception of two groups, they were largely infertile.

Number of worms per cocoon. At 20 °C the number of worms contained in each cocoon varied drastically with the day length at which the cocoon was laid (Table 1). The mean number of worms for all 58 fertile cocoons laid under long days was 2·25 (σ, ± 0·91) and for all 56 fertile cocoons laid under short days was 4·54 (σ, + 1·21). The probability that this difference arose by chance is much less than 0·001. Individuals which hatched from cocoons with few worms were rather larger than those hatched from cocoons with many worms. These larger individuals fissioned much earlier than small worms, sometimes as soon as a few days after hatching.

At 23 °C not enough fertile cocoons were laid to evaluate the effect of day-length. All 26 °C populations experienced one 6 °C drop in temperature. The cocoons laid by these groups did not show clear-cut differences.

Our experiments further confirm our previous results (Vowinckel, 1970b) that the daily illumination period decisively influences planarian reproductive physiology and that environmental temperature and photoperiod combine in the determination of the pathway reproduction is to take at any time. How then do these two factors exert their influence?

At the three temperature levels that were used we find completely different pathways of reproduction. Throughout the populations of one temperature level, however, reproductive behavior is fairly uniform.

At 20 °C cocoons were laid by nearly all populations, and asexual reproduction was limited to a minimum. While growth at this temperature continued due to high food intake, much of the gain seemed channelled into supernumerary reproductive structures. Jenkins (1970) described the occurrence of supernumerary gonopores in D. dorotocephala, and found that the presence of four to six gonopores interfered with reproductive success while two or three gonopores did not. In our stock there were very few individuals with four gonopores. It is conceivable, though, that supernumerary gonopores somewhat contributed to the number of infertile cocoons laid by our populations at 20 °C.

At 26 °C asexual reproduction continued toward exhaustion and sexual reprouction did not take place.

At 23 °C reproduction oscillated between development of sexual structures and fission. One gets the impression that both reproductive pathways fight for expression. In most populations sexual reproduction eventually resulted in some cocoon laying but these cocoons, with two small exceptions, were all infertile. This general infertility of cocoons at 23 °C may be due to a failure of sperm to survive at this temperature. This was suggested by Cowles (1965).

It is well known that temperature changes can have an inductive effect. For D. tigrina Vowinckel (1970 a) showed that temperature oscillations can be correlated with increases in the number of germ cells forming testicular primordia. This is in agreement with our present set of experiments in which the effect of relatively constant temperature levels, as we interpret it, is largely non-inductive, limiting and permissive.

The differences between populations at the same temperature level, on the other hand, can all be correlated with differences in illumination regimen.

The daily illumination period appears to play a regulating role in planarians much as it does in higher metazoa. That is to say, we witness the effects of a biological clock. In D. tigrina the photoperiod seems to determine the time of cocoon deposition. Maintenance under long photoperiods leads to later cocoon deposition than under short photoperiods. The natural generation length of D. tigrina, at our latitude, is approximately 365 days. Physiological activity in the cold winter months presumably comes to a near standstill leading among other things to an arrest of the biological clock. In the laboratory, in the absence of low temperatures, we therefore arrive at much shorter generation lengths than under natural conditions.

The photoperiod also influences the number of worms hatching from a cocoon. We do not know at what time this influence is exerted. Worms which experienced changes in photoperiod directly after hatching laid cocoons with a sibling number corresponding to the daylength which they had been shifted to. It would follow, therefore, that the number of offsprings contained in a cocoon was determined after the mother had hatched. One could speculate that the photoperiod influences the number of ova released from the ovary. Long days would reduce this number while short days would increase it. Long days, in general, seem to retard sexual development and encourage asexual reproduction.

Following a 6 °C drop in temperature this correlation between daylength and sibling number/cocoon is lost. We know already that germ cells respond to temperature oscillations with an increase in number. It seems plausible to suggest that temperature might also influence the number of worms hatching per cocoon and thus cause the loss of a simple correlation.

Under natural conditions cocoons are laid during the longest days of summer. After hatching individuals soon fission several times before sexual development commences. Cocoons laid under long-day conditions contain fewer worms of greater size, which fission earlier than short-day cocoons. Therefore, the fission period of sexually derived individuals would, under natural conditions, tend to occur earlier if cocoons are laid at the height of summer than if cocoons were laid in spring, that is under short-day conditions. Apart from this size factor long-day illumination by itself seems to promote faster and higher fission rates at 20 °C. The S-L shifted groups were laid under short days but they nevertheless fission earlier than the L-S populations. All long-day groups, finally, reach 100 % fission production while none of the short-day groups do so (Table 2).

We suggested that there exists during ontogeny a sensitive period during which the worm responds to daylength resulting in the determination of a corresponding generation length. This, presumably, would take place via the medium of some neurosecretory factor. Evidence of the existence of such factors was brought by Lender (1964) and Ude (1964). On the basis of our results we can, at present, only make some educated guesses as to what phase in the life cycle of D. tigrina is receptive to photoperiod induction. One of us (C.V.) suggested earlier (1970a) that it might be found during early development of the worm.

Worms were exposed to changes of photoperiod during their early development. This resulted without exception in a retardation of the onset of sexual reproduction in individuals which hatched from such cocoons. If we assume that this retardation represents interference with a mechanism which has been established or is in the process of being established we can consider several possibilities.

If induction took place directly after hatching, then all populations that were shifted before hatching should show the same generation length and fertility as the undisturbed groups of the daylength they were shifted to and the populations that were shifted 2 weeks after hatching should show signs of interference by retarded generation length and decreased fertility. According to our results such induction directly after hatching cannot have taken place. The populations that were shifted from long day to short day before hatching both have much longer generation lengths than the undisturbed short-period groups. The discrepancy is even larger with a shift in the opposite direction.

If induction had taken place later than 2 weeks after hatching, we would expect that all shifted groups should show the generation length of the daylength to which they had been shifted. Our results obviously do not agree with these premises either.

If induction took place during development in the cocoon we would expect the generation length of those groups that were shifted at laying to coincide with that of undisturbed groups of the photoperiod towards which they were shifted, and all later shifts to show interference. Again there is no evidence that this took place.

If, however, photoperiod acted on the maternal organism and the generation length was already determined before cocoons were laid we would expect to find evidence of interference in the generation length and fertility of all populations that were shifted. Our results obviously agree best with this assumption. In general, the later during ontogeny our shifts took place and the more they contradicted natural conditions the more severe was the interference. Possibly this means that at the onset of development this mechanism is still somewhat flexible to changes (especially around hatching time) while later on it appears to consolidate and loses its adaptability.

The populations derived from cocoons laid under natural conditions provide a convincing example of the importance of early development for photoperiod induction. After hatching these groups were exposed to widely different illumination régimes but responded with a very uniform generation length. The group that had no photoperiod interference during development reproduced earlier by about 2 weeks while the generation length of all populations shifted to incubators before hatching falls within 238 ± 6 days. Considered in combination with our other results this to us represents convincing evidence that the period which is sensitive to photo-induction of generation length is indeed found during the early development of D. tigrina.

It must be assumed that photoperiod control over reproduction has evolved most often in species or races whose environment permits reproduction only on a seasonal basis. Jenkins & Brown (1963) have described a race of D. doroto-cephala which lives in a constant temperature spring at 18-5 °C and produces cocoons permanently, apparently without seasonal variations. Obviously such an environment would favour the suppression of any seasonal timing of reproduction. The same may be true for a race of Cura foremanii which in our laboratory produces cocoons throughout the year.

This work was supported by a grant from the National Research Council of Canada.

Cowles
,
R. B.
(
1965
).
Hyperthermia, aspermia, mutation rates and evolution
.
Q. Rev. Biol.
40
,
341
367
.
Jenkins
,
M. M.
(
1970
).
Sexuality in Dugesia dorotocephala.
J. biol. Psychol.
12
,
71
80
.
Jenkins
,
M. M.
&
Brown
,
H. P.
(
1963
).
Cocoon production in Dugesia dorotocephala (Woodworth) 1897
.
Trans. Am. microsc. Soc.
82
,
167
177
.
Lender
,
T. H.
(
1964
).
Mise en évidence et rôle de la neurosécrétion chez les planaires d’eau douce (Turbellariés, Triclades)
.
Annls Endocr.
25
,
61
65
.
Siegel
,
S.
(
1956
).
Nonparametric Statistics.
New York
:
McGraw-Hill Co
.
Ude
,
J.
(
1964
).
Untersuchungen zur Neurosekretion bei Dendrocoelum lacteum Oerst. (Platyhelm. Turbell
.).
Z. wiss. Zool.
170
,
223
255
.
Vowinckel
,
C.
(
1968
).
The Induction of Sexual Development in the Planarian Dugesia tigrina (Girard), Doctoral dissertation, McGill University, Montreal, 104 pp
.
Vowinckel
,
C.
(
1970a
).
The stimulation of germ cell proliferation in Dugesia tigrina (Girard)
.
J. Embryol. exp. Morph.
23
,
407
418
.
Vowinckel
,
C.
(
1970b
).
The role of illumination and temperature in the control of sexual reproduction in the planarian Dugesia tigrina (Girard)
.
Biol. Bull. mar. biol. Lab., Woods Hole
138
,
77
87
.
Vowinckel
,
C.
,
Wolfson
,
N.
&
Marsden
,
J. R.
(
1970
).
A crowding factor causing spontaneous decapitation in cultures of the planarian Dugesia tigrina.
Can. J. Zool.
48
,
1059
1062
.