Differential Impacts of the Head on Platynereis dumerilii Peripheral Circadian Rhythms

Animals were maintained under controlled temperature and on 16:8 h light-dark (LD) or dark-dark (DD) cycles as previously described (Schenk et al., 2019). Sampling points are presented either as zeitgeber time (ZT) for LD conditions or circadian time (CT) for constant darkness (DD). ZT0 is defined as the time of light ON, CT0 defined as the time the light would go ON under DD conditions. All animal work was conducted according to Austrian and European guidelines for animal research.

Animals were anesthetized by adding a few drops of 1 M MgCl₂ in the seawater until they stopped moving. They were carefully placed on a microscope slide under a binocular dissecting microscope, decapitated, and transferred to fresh seawater again. Decapitation was done with sterile surgical blades cutting on the first segments below the pharyngeal region in order to collect only the posterior part of the body (i.e., the trunk). They were consistently performed between ZT 13.5 and ZT14. For re-entrainment, ZT refers to the actual ZT of the new LD regime. For head samples, the region containing the pharynx and the posterior end of the head was removed (see Zantke et al., 2013).

Groups of animals used for DD experiments were collected between ZT 13.5 and ZT14 on individual plastic cups with seawater, wrapped in aluminum paper and placed inside a custom-made dark box in a temperature-controlled environment. After the proper acclimation time (1–5 days depending on the experiment) a plastic cup was taken out at a given CT and animals were used for RNA extraction/RT-qPCR or pigmentation quantification. This method ensured the avoidance of risking a premature exposure to light to the remaining samples in the dark box. Further animal handling (e.g., for decapitation or chromatophore visualization) was done under dim light and as fast as possible.

When testing for circadian re-entrainment (i.e., an inverted light cycle) under white, blue, or red light, animals were exposed to the new conditions for 7 days before sampling. Light spectra and intensities of white, blue, and red LEDs (ProfiLuxSimu-L from GHL advanced technology gmbh, Germany) used for circadian re-entrainment were measured using an ILT950 spectrometer (International Light Technologies Inc., Peabody, United States). Special care was taken to account for the standard conditions, where the worms were housed, i.e., 22 cm away from light source and with a transparent plastic lid positioned between the spectrometer and the light source. Measured light intensities for white, red, and blue lights were 8.2 × 10¹³ photons/cm²/s, 3.8 × 10¹³ photons/cm²/s, and 2.4 × 10¹³ photons/cm²/s, respectively (for spectra, see Supplementary Figure 1).

Total RNA was extracted from heads or trunks (i.e., decapitated animals) using the RNeasy Mini Kit (QIAGEN). In the case of heads, each biological replica consisted of 4–5 heads (to obtain sufficient amounts of RNA), while for trunks each replica consisted of a single decapitated animal. Reverse transcription was carried out using 0.4 μg of total RNA as template (QuantiTect Reverse Transcription kit, QIAGEN). RT-qPCR analyses were performed using a Step-One-Plus cycler. The expression of each test gene was normalized by the amount of the internal control gene cdc5. The relative expression was calculated using the formula 1/2^ΔCt. Primers and PCR program used are listed in Zantke et al. (2013).

Three consecutive segments located toward the middle of the body were selected on each animal to evaluate changes in chromatophore size. In order to precisely re-identify the same segments over the course of the experiments, animals were anesthetized with MgCl₂ and a parapodium, two segments away from the region of interest, was removed with a sterile surgical blade. When required, one animal at a time was placed on a glass cover without water and extended carefully. An epifluorescence stereoscope (Zeiss Lumar) with a 488 nm laser and a FITC filter was used to take pictures making sure to always use the same magnification across animals and sampling points. Animals were placed again in seawater until the next sampling point. Image analysis was done using Adobe Photoshop. On fluorescent images, the RGB channels red and blue were lowered to zero, and the three segments of interest were extracted by erasing the unwanted area (chromatophores on each segment have a specific pattern, which makes their visual identification easier). A new layer was generated by using the magic wand tool to single out the bright green chromatophores from the background fluorescence. Using the image’s histogram, the number of colored pixels was used as a proxy of chromatophore size. Animals had between 20 and 40 chromatophores along the three segments, but no effort was done to quantify size of individual chromatophores. Instead, the sum of all the chromatophores of interest were used as chromatophore size value at each time point. Absolute pixel number was expressed as a percentage of the maximum value for each animal across all time points of the experiment [i.e., x_i/(Max{x_i,…,x_j}^*100)]. Average among biological replica and SEM were further calculated.

Immature worms of comparable size were starved for 3 days before the start of the assay. After decapitation, worms were placed in individual hemispherical concave wells (diameter = 35 mm, depth = 15 mm) of a custom-made 36-well clear plastic plate [as described in Ayers et al. (2018)]. Intact worms were also treated with MgCl₂ for 5 min prior to locomotor recording to ensure proper comparisons to decapitated worms. Video recording of worm’s behavior over several days was accomplished as described previously (Zantke et al., 2013), using an infrared (λ = 990 nm) LED array (Roschwege GmbH) illuminating the behavioral chamber and an infrared high-pass filter restricting the video camera. Worms were recorded over 4 LD cycles (16 h light/8 h darkness), followed by 3 days under constant darkness. White light was generated by custom made LEDs (Marine Breeding Systems, St. Gallen, Switzerland) with the intensity set to 5,2 × 10¹⁴ photons/cm²/s at the place were worms were housed (for light spectrum see Supplementary Figure 1). Trajectories of locomotor activity of individual worms were deduced from the video recordings by an automated tracking software developed by LoopBio gmbh¹ (Vinoth Babu Veedin Rajan et al., unpublished). Locomotor activity trajectories reflect the distance moved of each worm’s center point across 6 min time bins. Activity data was plotted as double-plotted actograms using the ActogramJ plugin for Fiji (Schmid et al., 2011). For primary data see Supplementary Data 2.

The main statistical analyses were performed using either the data analysis plug-in in Microsoft Excel using an alpha value of 0.05 (molecular and chromatophore data) or GraphPad Prism (locomotor activity data). For changes in chromatophore size, each two-sample two-tailed student’s t-test was preceded by an F-test to check if the variances of the two groups were equal or not. In the cases were the same animals were used over time (i.e., repeated measures), a paired two-tailed Student’s t-test was used. To test if transcriptional changes in gene expression over time oscillated to a statistically significant difference (i.e., the curve was not statistically linear), fold change data was analyzed for each sampling point using single factor ANOVA. In order to ease the logistic process of analyzing a considerable amount of data, RNA samples and chromatophore size images were not analyzed blindly but chronologically as experiments were being performed. Two-way ANOVA and post hoc tests were performed in R v3.5.1 (R Core Team, 2018) with the package emmeans. All data sets were checked if the assumption of homogeneity of variances and normality were met. In the case of repeated measures ANOVA, GraphPad Prism software was used; sphericity was not assumed and the method of Greenhouse and Geisser was used to adjust the results (see Supplementary Data 1 for details on two-way ANOVA analyses). Statistical differences in locomotor activity across treatments were estimated using repeated measures ANOVA (GraphPad Prism) followed by Sidak’s multiple comparison test. To identify the free-running period length of intact worms under DD conditions Lomb-Scargle periodogram analysis was done using the ActogramJ plugin for Fiji (Schmid et al., 2011).

All sequence resources referred to here were already published previously and submitted to public databases. All other data that support the findings of this study are available from the corresponding author upon reasonable request.

As a starting point, we aimed to test for the presence of peripheral clocks in the body of P. dumerilii adults. Based on the previously described and tested molecular components of the core circadian oscillator in P. dumerilii heads (Zantke et al., 2013), we analyzed the daily transcriptional fluctuation of bmal, period and tr-cry as representative components of the circadian clock by RT-qPCR (Figure 1). Daily fluctuation of bmal, period, tr-cry, timeless, clock, and pdp1 transcripts in the trunk are overall similar to those in the head [Figures 1A,B, for timeless, clock, and pdp1, see Supplementary Figures 2A–C, compare also with (Zantke et al., 2013)], although overall relative mRNA levels were always significantly higher in heads than in trunks. In the case of period, there was also a significant difference in overall transcript levels between trunks under LD and DD conditions, with higher levels at LD (pairwise comparisons of estimated marginal means, Tukey adjusted with alpha = 0.05) (see Supplementary Data 1 referring to Figure 1). Results were similar irrespective of the segment positions used for analyses, i.e., whole trunk (Figure 1), last 5–7 segments of the body or the adjacent 5–7 segments toward the anterior part of the animal (Supplementary Figures 2D–F) produced consistent results. Pdu-tr-cry appears to be the gene that deviates most in this comparison. We attribute this difference to a higher variability of the transcript level synchronization in the trunk given that in most cases it corresponds to the oscillations observed in the head. Alternatively, the presence of the head leads to a shift in the tr-cry curve maxima, as the curves for tr-cry consistently differ when the head is present (compare Figures 1A,B, 2A,B, 3C and Supplementary Figure 2F, and see also Zantke et al., 2013 for heads, Tosches et al., 2014 for whole larvae). A two-way ANOVA on ZT and condition (i.e., LD Heads and LD Trunks) was carried out to further analyze tr-cry transcript changes. There was a statistically significant effect of ZT, condition and, most importantly, interaction between the ZT and the condition on relative mRNA levels [F(5,95) = 5.3837, p = 0.0002]. A post hoc pairwise Tukey analysis for the interaction shows that the low relative mRNA levels at ZT5 on trunks under LD conditions is the main difference between these two conditions (see Supplementary Data 1).

We next tested if the peripheral circadian clock transcript oscillations would continue synchronously under constant darkness for 3 days. All tested transcripts dampened strongly, with tr-cry still exhibiting weak significant changes across CT (Figure 1C).

To analyze if the peripheral circadian clock transcript changes in the worm’s body were dependent on the circadian clock in its head, we took advantage of the fact that bodies of decapitated animals survive in seawater for about 2 weeks (Hofmann, 1976). Adult animals were decapitated and placed under standard LD cycles before RNA extraction. Relative transcript levels of bmal, period, and tr-cry exhibited overall similar levels and continued a clear diel cycling of expression for at least 3 days in trunks of decapitated animals (Figure 2A), indicating that the peripheral circadian clock gene expression continues to synchronously run even in the absence of the brain circadian clock. We then tested core circadian clock transcript changes in trunks of decapitated animals under constant darkness. On the first day of DD (DD1) trunks continue to show a significant oscillation over time in all three genes (Figure 2B). However, after 3 days in complete darkness (DD3), relative mRNA levels dampen and no significant changes were detectable for any of the three genes (Figure 2C). We also tested for overall relative mRNA levels of expression of each gene under each condition (pairwise comparisons of estimated marginal means, Tukey adjusted with alpha = 0.05, see Supplementary Data 1 for Figure 2). In the trunk bmal overall levels were higher in LD than in DD1 and DD3 conditions (which in turn were statistically similar among them), while overall period levels were higher in DD1 compared to DD3 and LD conditions (statistically similar among them). For tr-cry, although LD is also clearly still higher, differences in overall levels on all three conditions are statistically significant (see Supplementary Data 1 referring to Figure 2). These differences possibly reflect differences in the transcriptional activation and repression during the different light conditions, e.g., it is noteworthy that under DD1 levels of the transcriptional suppressor per get significantly higher, while the transcript levels of the corresponding “downstream” transcriptional activator bmal are lowered under DD1, as would be expected if more per is present and translated to protein.

By placing decapitated animals on an inverted light cycle, we next tested for the capacity of peripheral clocks to be re-synchronized by light, again using the transcript oscillations of bmal, per, and tr-cry as readout. Cycling of bmal and per was re-entrained to the inverted cycle when exposed to white, red or blue light (Figures 3A,B, for spectra and intensity see Supplementary Figure 1). tr-cry transcript oscillations differed from this, in that white and blue light could re-entrain its peripheral oscillations as in the case of bmal and per, whereas red light did not (Figure 3C). To better understand these findings, a two-way ANOVA to simultaneously examine the effect of light treatment (i.e., light color = wavelength) and ZT was performed for each gene (see Supplementary Data 1). In the case of bmal, ZT [F(5,146) = 12.6907, p = 3.0399e^–10] had a significant effect; with ZT5 and ZT21 having the same levels of expression. A similar result was found for per, but with the important addition of a significant interaction between ZT and the light condition (F(10,144) = 2.7811, p = 0.0036), indicating that the relationship of ZT and relative mRNA levels is dependent on the type of light. A post hoc pairwise Tukey analysis for the interaction indicates that such dependence is due to a specific variation on relative mRNA levels of per on tails under red light at ZT14 and ZT21. In the case of tr-cry, the situation is more complex. The analysis shows an effect of ZT, Light and a significant interaction between the two [F(10,144) = 2.1109, p = 0.0271] on relative mRNA levels. The post hoc pairwise analysis of light treatments (i.e., colors) over each ZT shows statistical differences for red light at ZT14 and ZT21, with significantly higher relative mRNA values compared to white or blue light at the same ZTs (Supplementary Data 1 referring to Figure 3). These punctual higher relative mRNA levels are the clearest evidence for the mentioned lack of re-entrainment of tr-cry peripheral oscillations by red light. Overall, these results demonstrate that circadian clock transcripts in peripheral tissues can directly respond to changes in the light cycle, independently of the head.

In order to assess how our findings on core circadian clock transcripts oscillations might relate to physiology and behavior, we investigated possible outputs of peripheral circadian clocks, starting with changes in chromatophore size in the trunk. Chromatophores are located along the dorsal part of the segmented body of P. dumerilii. Based on light microscopy analyses it had previously been shown that the worm’s chromatophores exhibit a segment-autonomous, diel contraction-expansion rhythm with increasing size during the day and decreasing size at night (Fischer, 1965; Röseler, 1969, 1970).

We first aimed at identifying a possibility to automatize the recording of chromatophore changes. We found that chromatophores exhibit a well-detectable autofluorescence under 488 nm light (Figures 4A,B), which can be used for automatic detection by any image software. In order to characterize the pattern of contraction/expansion of the chromatophores, animals were photographed every 3 h over the course of 24 h using a fluorescence microscope. We found a clear circadian pattern with higher chromatophore expansion between ZT2 and ZT11 (Figure 4C), and a sharp drop on chromatophore size from ZT11 to ZT14, before lights go off and the subjective night period starts, already suggestive of an autonomous clock-driven process and not a direct light response, again consistent with previous observations (Hempelmann, 1939; Fischer, 1965).

We next focused on sampling points corresponding to ZT/CT2 and ZT/CT14, during which a ∼60% drop in chromatophore size was evident (Figures 4C, 5A,B), and used these two time points as a reference for the study of circadian cycling of pigmentation over multiple days. Chromatophores expansion/contraction continued to cycle in animals placed under constant darkness for five consecutive days (Figure 5C).

For evidence that the light used for measuring chromophore size is not causing re-entrainment in this case, see below.

In order to assess if the regulation of the cycling on chromatophore size is governed by peripheral clocks, decapitated animals were used. The same animals were photographed at ZT2 and ZT14 prior to decapitation as a starting reference point. Following decapitation, worms were placed in constant darkness for 4 days, and subsequently exposed again to a normal LD cycle for additional 3 days. Chromatophores sizes were registered along the experiment from day 0 to day 4, and again on day 7 to test for possible re-entrainment. Upon decapitation, animals under DD conditions initially continue to exhibit clear rhythms of chromatophore size changes (Figure 5D, for individual replicas see Supplementary Figure 3). Starting with the second day in DD, the rhythm will, however, start to dampen and become statistically non-significant by day 4 (Figure 5D). Subsequent exposure to a normal LD cycle did not lead to a re-synchronization of the chromatophore rhythm (Figure 5D). Consistently, cycling of chromatophore size does not get re-entrained on decapitated animals under an inverted LD cycle applied for 5 days (Figure 5E). A two-way repeated measurement ANOVA and post hoc pairwise analyses, confirmed a significant interaction between CT (i.e., time of subjective day) and the different days in DD [F(3,15) = 13.34, p = 0.0050], indicating that the difference between CT2 and CT14 gets significantly smaller over time from DD1 to DD4 (see Supplementary Data 1 referring to Figure 3).

In order to rule out that we may have missed phase-shifts on decapitated animals due to the exposure to the 488 nm light during the measurement procedure (due to too low sampling frequency), we also performed a more densely spaced 24 h sampling on day 5 in LD (post decapitation). This experiment confirmed our interpretation of a dampening of the chromatophore size rhythm and inability to re-entrain in the absence of the head (Figure 5F). All together, these results suggest a circadian pattern of chromatophore size governed by a peripheral clock, which, however, requires the head to maintain extended synchronization and for re-entrainment by light.

We next turned to locomotor activity as a read-out for circadian clock activity. We have previously shown that P. dumerilii during new moon (circalunar phase of its circalunar clock) exhibits nocturnal circadian locomotor activity, which free-runs under constant darkness for at least 3 days (Zantke et al., 2013, 2014). We meanwhile established an automated worm locomotor behavioral tracking system, which measures worm activity as distance moved of the worm’s center point within 6 min time bins [Ayers et al., 2018, LoopBio (see footnote 1), Vinoth Babu Veedin Rajan unpublished]. It should be noted that this new type of analysis measures relative distance moved, compared to the binary (movement: yes or no) manual scoring done by Zantke et al. (2013). Intact worms showed a significant circadian rhythmic locomotor activity under LD and DD conditions (Figures 6A,B, for individual actograms see Supplementary Figure 4). In contrast decapitated worms exhibit an overall severe reduction in movement and rhythmicity (Figures 6C,D, for individual actograms, see Supplementary Figures 5, 6 and Supplementary Movies 1–6. Please note that the maximal values of the Y-axis for decapitated animals is 5–10 smaller than for non-decapitated controls in order to visibly display movement). Overall, these data suggest that lack of signals from the head lead to a general suppression of movement and lack of circadian information or synchronization signals for the locomotor output. Headless worms are not generally paralyzed, as they can show spontaneous bursts of movement (Supplementary Figures 5, 6 and Supplementary Movies 1–6) and- depending on their stage at decapitation- can proceed to maturation and the associated behavioral changes (Hauenschild, 1966). Similar to the rhythmic transcript oscillations of the core circadian clock genes in the trunk, acute light functions as a synchronization cue to the periphery of headless worms, but without head or light stimuli a circadian locomotor pattern cannot be maintained by the trunk alone in the majority of the worms.

Here, we examine peripheral circadian clock transcript changes and diel rhythms in chromatophore size and locomotor behavior in P. dumerilii in the absence and presence of its head. An endogenous circadian rhythm driving body pigmentation change in P. dumerilii had been previously proposed based on photographic recordings of isolated groups of body segments during the middle of the day and the night (Röseler, 1969, 1970). With the exception of gills in oysters (Payton et al., 2017), there is no information on fluctuation of circadian clock genes on peripheral tissues or appendages in marine organisms. We document the expression of core circadian clock genes in the periphery of P. dumerilii, arguing in favor of a functional circadian peripheral clock and opening an avenue to study the molecular mechanisms of peripheral clocks in marine invertebrates. We show that light/dark cycles can (at least in part) substitute for the head as major synchronizer for continuous peripheral core circadian clock transcript oscillations. We confirm previous work on chromatophore rhythms, which – in contrast to locomotion and transcript oscillations- exhibit free-running rhythms in trunks of headless worms.

As previously reported for the central clock (Zantke et al., 2013), period and bmal transcripts cycle in antiphase from each other, while tr-cry transcripts are neither directly in-phase nor in anti-phase with any of them. Our results on trunks show that cycling of bmal and period continues under LD conditions, but not under DD independently of the head being present or absent, and can be entrained to an inverted LD cycle on decapitated animals. These results suggest that the peripheral clocks are light dependent (through a yet to be identified set of photoreceptors in the trunk) and either stop, or more likely, desynchronize in the different peripheral cells and tissues (at least as fast as 3 days in DD) in the absence of the head. Consistent with endogenous oscillators desynchronizing over time this desynchronization does not happen immediately upon placement in DD conditions (i.e., DD1). A system with peripheral clocks independent from the central circadian clocks in the brain and entrained directly by environmental signals is reminiscent of Drosophila melanogaster (Tomioka et al., 2012). In that sense the peripheral clock in P. dumerilii resembles that of insects, and light reception in the trunk likely occurs via Cryptochromes and/or Opsins. Candidates include Go-Opsin1 and rOpsin1 (Backfisch et al., 2013; Ayers et al., 2018).

Light exhibits different effects on the different readouts. In the cases of transcript oscillations and locomotor activity the head is not required for its impact, suggesting that peripheral photoreceptors mediate this signal. Interestingly, different wavelengths appear to have differential peripheral effects on transcript oscillations (red light being able to re-entrain per and bmal trunk oscillations, but not tr-cry), which already indicates the involvement of more than one photoreceptor. It will be interesting for future studies to understand why tr-cry transcripts behave differently from bmal and per transcripts under different light conditions. These differences could be the result of tr-cry only being expressed in a subset of tissues/cell types that do not desynchronize as quickly.

As in the case of insects (Tomioka et al., 2012) and mammals (Reppert and Weaver, 2002), questions regarding how peripheral clocks are entrained and the actual mechanisms that peripheral clock use to drive transcriptional changes on various tissues are still questions to be answered in P. dumerilii, as is the case for the specific functions of the peripheral clocks.

It should be noted that our experiments were performed on the complete trunk (or at least several segments), which also leaves open questions regarding the peripheral circadian cycling and their synchronizations in specific segmentally repeated organs and tissues.

Daily changes in chromatophore size on P. dumerilii can be easily used as a visual read out of the circadian clock, adding up to its locomotor activity, circalunar spawning, and clock-related genes as means to study chronobiology in this model system.

Remarkably, the first studies on the cyclic changes of body pigmentation in P. dumerilii date back to Hempelmann (1939). Based on this and further classical work, these changes in body pigmentation were already attributed to an endogenous circadian rhythm present in each segment of the worm’s body (Fischer, 1965; Röseler, 1969, 1970), pointing at the existence of peripheral circadian clocks long before their molecular mechanism had been unraveled and circadian peripheral oscillations were proven to exist in the peripheral tissues in mammals (Tosini and Menaker, 1996; Balsalobre et al., 1998; Schibler et al., 2003; Dibner et al., 2010). Our analyses support the classical studies on Platynereis, in that chromatophore size in the body of P. dumerilii is higher during day time, with a major drop before the night begins; which argues in favor of a clock-driven manifestation and not a direct light response. The average magnitude of this change corresponds with previous quantitative reports by Fischer (1965). We report that individuals placed in DD for 5 days still exhibit a significant daily difference on chromatophore size, although its maximum value decreases by 25% compared to the initial LD conditions. It has been reported that such cycling starts to fade on individual chromatophores after 7 days in DD (Fischer, 1965), but we did not test for the long term stability of the cycle for individual chromatophores. Noticeably, a similar circadian cycling of chromatophore size on DD and inverted LD cycles has been shown for the marine isopod E. pulchra (Zhang et al., 2013), posing the questions if this regulation is similar.

Changes in chromatophore size with a circadian cycling is commonly seen in crustaceans (Fingerman, 1955; Fingerman and Yamamoto, 1967; Fingerman et al., 1969; Zhang et al., 2013). There is usually an increase in size during the day thought to be related to UV protection (Darnell, 2012), but an inverted pattern of bigger chromatophore size during the night, to possibly enable individuals to camouflage with the substrate, can also be found (Stevens et al., 2013). The fact that P. dumerilii is mostly active during the night (Zantke et al., 2013), when pigmentation is lower, does not argue in favor of a camouflage role; especially since pigmentation does not respond to changes in background brightness (Fischer, 1965). The most parsimonious option is therefore, a role of pigmentation in protecting the animal from UV light. It should be noted, however, that it has also been often proposed, but not tested, that circadian changes in pigmentation might be a mechanism related to energy saving (Stevens, 2016).

Remarkably, in our hands the re-entrainment of chromatophore rhythms by light requires the presence of the head. This might be either because the required photoreceptor(s) are located in the head or because hormonal feedback signals, such as Pdu-PDF (pigment dispersing factor) (Shahidi et al., 2015), are required for the synchronization process. It is likely that pigmentation in P. dumerilii is controlled by a hormonal process, as in some crustaceans (Fingerman and Yamamoto, 1967; Fingerman et al., 1969; Nery et al., 1999), which in turn is governed by the central clock. The hormonal nature of the cycling on chromatophore size has been further supported by the immediate reaction of chromatophores, present on isolated skin tissue, when coelomic fluid from P. dumerilii during a given time point (e.g., day or night) is added (Röseler, 1970).

Finally, while we overall confirm previous work on the chromatophore rhythms in the trunk of P. dumerilii, there is one clear difference between our findings and that of Röseler (1970). Her work shows that chromatophore rhythms can still be re-entrained by white light even in the absence of the prostomium (head), whereas in our hands decapitated animals do not re-entrain their chromatophore rhythm in response to white light. We identified two main reasons that might explain this discrepancy. The materials and methods of her paper do not state the light intensity and spectrum. It is thus possible that this strongly deviates from our conditions. The other difference is the extent of head removal. In our study we removed the head including the jaw piece, whereas Röseler (1970) specifies prostomium-removal, which implies that her worms still possessed the jaw and the surrounding tissue. This region possesses multiple neurons and neurosecretory cells, which could be important for proper re-entrainment. Further work will certainly help to disentangle these differences.