Live to Die Another Day: Regeneration in Diopatra aciculata Knox and Cameron, 1971 (Annelida: Onuphidae) Collected as Bait in Knysna Estuary, South Africa

Simple Summary The estuarine moonshine worm, Diopatra aciculata, is used extensively as bait in the Knysna Estuary in South Africa. During collection, the worm frequently breaks into multiple pieces. If discarded or unused pieces can regenerate to form separate individuals, the population may be maintained, or even increase, despite harvesting. This study investigated bait collecting habits of local fishermen and the natural incidence of regeneration in D. aciculata. Fishermen usually removed only part of the worm, leaving its tail in the tube and more than half the fishermen return up to 50% of bait collected to the estuary. Naturally occurring D. aciculata can regenerate missing anterior and posterior chaetigers, but only if amputation occurs before the 17th or after the 21st segment. Most unused fragments are probably too small to recover from damage inflicted during bait collection, so regeneration is unlikely to cause population expansion despite harvesting. However, some fishermen do move bait from the estuary. Range expansion can therefore occur if large fragments discarded at fishing sites in other estuaries do regenerate, forming new populations. Abstract Regeneration is critical for survivorship after injury, sublethal predation, and asexual reproduction; it allows individuals to recover, potentially enabling populations of bait species to overcome the effects of bait collection through incidental asexual reproduction. Opportunities for regeneration are created when worms break during collection (which happens more often than not) and are thrown back into the estuary. Additionally, the trade and movement of bait could result in the range expansion of invasive species. This study investigated bait collection habits of local fishermen and the in situ incidence of regeneration in the estuarine moonshine worm, Diopatra aciculata. The evidence shows that this species is capable of anterior and posterior regeneration. The disproportionately small percentage of worms that seem to be recovering from the degree of damage that may be inflicted during bait collection suggests that regeneration may not help worms to withstand the effects of bait collection. However, the continuous movement and discarding of even small numbers of bait in other estuaries can lead to range expansion through incremental build-up, forming new populations, if these fragments are large enough to regenerate.

Extensive posterior regeneration linked to density-dependent aggression has been reported for D. aciculata in farmed populations [6], at a proportion that far exceeds that observed in any other Diopatra species [36]. However, whether D. aciculata can regenerate anteriorly, and to what extent, is still unknown. Diopatra aciculata is morphologically very similar and very closely related to D. neapolitana [20] and it is therefore possible that they would show similar regenerative potential. The branchiae in D. aciculata start from the fourth or fifth chaetiger and extend for 20 to 40 chaetigers [21], whereas the branchiae in D. neapolitana start at chaetiger three or four and extend for approximately 45 to 55 chaetigers [36]. Diopatra neapolitana can regenerate anteriorly if the amputation site was around the 15th chaetiger, whereas amputation at the 20th chaetiger led to the death of both parts of the worm. Amputation after the 25th chaetiger, however, allowed for posterior regeneration [36]. If the number of chaetigers where regeneration is possible is proportionate to the total number of branchiate chaetigers, and if these are similar in D. aciculata, we can expect anterior regeneration if amputation occurs around the 5th to 13th chaetigers and posterior regeneration if amputation occurs after the 9th to 22nd branchial chaetigers in this species.
Worms are often damaged during bait collection, creating fragments that could regenerate in several ways: only a fraction of a specimen is usually removed from the tube, leaving the posterior fragment in situ [46]; the individual may break into pieces after being removed from its tube [47]; and leftover bait is often discarded by fishermen and bait collectors [21,31,48]. If these fragments can regenerate, collecting and discarding bait has the potential to greatly affect population growth. Discarded pieces can regenerate, allowing populations to be maintained while incidental asexual reproduction can even allow for population growth.
The Knysna Estuary is situated within the Garden Route National Park and is managed by the South African National Parks (SANParks), who consequently conduct regular surveys of fishing and baiting activity in the area [49]. Over the period of January to December 2021, they determined that on average 45 people fish per a day during the week and 74 people fish per a day over weekends. This amounted to a total of 19,954 fishing days for that year [49]. Furthermore, they also made 494 observations of baiting and found that 12% were most likely collecting Diopatra. Finally, they report that recreational and subsistence fishers collect worms, while subsistence fishers may also sell worms to recreational fishers, including tourists, and that the bait sold by subsistence fishers are often used by recreational fishers to catch fish in areas outside of Knysna [50]. The ability to regenerate may also facilitate dispersal, but while there is evidence that the movement of bait species for trade has been implicated in the spread of invasive species [51], few studies have considered the effects of bait collection and the intraregional movement of this bait species.
This study investigates (1) the incidence of regeneration in Diopatra aciculata and (2) bait collecting behaviour to explore the potential for regeneration to facilitate population maintenance or expansion despite harvesting and if it could enable the dispersal and range expansion within Knysna Estuary and to other estuaries.

Materials and Methods
The Knysna Estuary on the south coast of South Africa (Figure 1) is an S-shaped estuarine bay approximately 20 km long, with a tidal flow and extensive intertidal flats [52]. The system covers an area of 10 km 2 at low tide and 16 km 2 at high tide with water supplied from the Knysna river, several smaller northern and eastern tributaries, and the permanently open mouth [53]. The estuary can be divided into three sections: the marine-dominated and strongly tidal lower estuary or "embayment" from the Western and Eastern Heads to the railway bridge; the marine-dominated middle estuary, from the railway bridge to the road bridge, dominated by warm water with strong salinity gradients; and the typically estuarine upper estuary, upstream of the road bridge, which is influenced by fluvial flow [54].
The system covers an area of 10 km 2 at low tide and 16 km 2 at high tide with water supplied from the Knysna river, several smaller northern and eastern tributaries, and the permanently open mouth [53]. The estuary can be divided into three sections: the marinedominated and strongly tidal lower estuary or "embayment" from the Western and Eastern Heads to the railway bridge; the marine-dominated middle estuary, from the railway bridge to the road bridge, dominated by warm water with strong salinity gradients; and the typically estuarine upper estuary, upstream of the road bridge, which is influenced by fluvial flow [54].

In Situ Regeneration
From January 2021 to June 2022, approximately forty specimens were collected monthly from Bollard Bay and The Point ( Figure 1). Individuals were collected using a 1 m length of piano wire ( [20] Supplementary Material Video S1), and taken to the laboratory for further analysis. Each individual was examined for signs of regeneration and classified as showing no regeneration, regenerating anteriorly, regenerating posteriorly or regenerating in both directions (bidirectional regeneration). When regeneration was present, the number of original branchiate chaetigers present, the number of chaetigers

In Situ Regeneration
From January 2021 to June 2022, approximately forty specimens were collected monthly from Bollard Bay and The Point ( Figure 1). Individuals were collected using a 1 m length of piano wire ( [20] Supplementary Material Video S1), and taken to the laboratory for further analysis. Each individual was examined for signs of regeneration and classified as showing no regeneration, regenerating anteriorly, regenerating posteriorly or regenerating in both directions (bidirectional regeneration). When regeneration was present, the number of original branchiate chaetigers present, the number of chaetigers regenerating anteriorly (excluding the prostomium and peristomium), and the number of chaetigers regenerating posteriorly were recorded. Since a fixed number of chaetigers will appear simultaneously during anterior regeneration [39], the number of chaetigers regenerating can inform the total number of branchiate chaetigers that were originally present and the extent of the damage being repaired. During posterior regeneration, however, chaetigers are added individually after a new posterior growth zone is established [40]; consequently branchiate chaetigers are only replaced near the completion of regeneration. The number of original branchiate chaetigers present therefore cannot be used to determine the exact number of chaetigers lost posteriorly but can be used to estimate the extent of damage to the individuals. Images of regeneration were taken on a Leica Stereomicroscope (Leica microsystems, Wetzlar, Germany; model number: Leica mz7.5) fitted with a Leica microscope camera (Leica microsystems, Wetzlar, Germany; model number: Leica EC3) and an Olympus Targus 5.

Interviews
Bait collectors and fishermen were interviewed throughout the estuary. The interviews were conducted from approximately two hours before low tide to two hours after low tide between June and December 2021. Interviews started at the upper reaches of the estuary, at the road bridge, and concluded near the estuary mouth, at Bollard Bay ( Figure 1). Bait collectors and fishermen were identified, approached, and invited to complete the questionnaire. Verbal consent was requested before questioning commenced (human ethical clearance number: REC-2021-19365). The questionnaire was designed to determine the bait preferences, collection practices and post-collection habits of the local fishermen as set out below ( Figure 2; Supplementary Video S1):

1.
To identify the fishermen who use D. aciculata, fishermen were asked to list their preferred bait species; only responses from those that selected D. aciculata were retained for analysis ( Figure 2, Q1).

2.
To estimate the magnitude of potential for the regeneration of D. aciculata, fishermen were asked how many worms they collected ( Figure 2, Q2).

3.
To assess if regeneration could lead to dispersal, respondents were asked if they moved bait within and out of the Knysna Estuary ( Figure 2, Q3a), as this creates an opportunity for anthropogenic dispersal. Secondly, respondents were asked if they bought Diopatra (Figure 2, Q3b), because recreational fishermen tend to purchase bait from subsistence fishermen. As recreational fishermen tend to fish in areas away from the subsistence fisherman (i.e., from whom bait is purchased), the likelihood of anthropogenic dispersal also increases if Diopatra are purchased as bait [49]. Furthermore, many recreational fishermen fish outside of Knysna [49].

4.
To assess the extent to which discarding unused bait could contribute to dispersal and to maintaining population size despite harvesting, fishermen were asked if they had bait left over and if yes, how the leftover bait was processed or discarded ( Figure 2, Q4).
In the latter instance, we only considered the discarding of fresh, unprocessed, bait. If large enough pieces of Diopatra are thrown back (size gleaned from observational data), a potential for regeneration is created. Once the worm fragments settle and regenerate fully, naturalisation is possible.

5.
The fishermen were asked which portion of the worm they preferred as bait (head, middle, tail, or whole worm), together with the frequency with which D. aciculata broke during collection (never 0%, rarely 0-33%, sometimes 33-66%, usually 66-99%, always 100%) ( Figure 2, Q5a and 5b). This information was used in conjunction with observations of in situ regeneration to explore if broken pieces of worm that are left behind during bait collection could regenerate and contribute to population growth or maintenance. The assumption was that if fishermen predominantly collected the portion of the worm that they preferred to use as bait, this section would predominantly be leftover and discarded, and these sections would therefore have the greatest potential to survive and, if large enough, regenerate. Additionally, the section of the worm left in the tube (i.e., usually the posterior) could also regenerate if large enough. If both anterior and posterior regeneration is possible, both portions can regenerate leading to incidental asexual reproduction.
tion of the worm left in the tube (i.e., usually the posterior) could also regenerate if large enough. If both anterior and posterior regeneration is possible, both portions can regenerate leading to incidental asexual reproduction.
Responses to the questionnaire were used in conjunction with observations of regeneration and data supplied by SANParks to explore if bait collection and regeneration can facilitate the persistence and anthropogenic dispersal of the species, as set out in Figure 2.

In Situ Regeneration
To calculate the proportion of branchiae that need to be intact for anterior regeneration, the following equation was used: % branchiae intact n origional branchiate cheatigers n regenerating branchiate cheatigers n original branchiate chaetigers 100 The percentage of original branchiate chaetigers needed for anterior regeneration was divided into 10 chaetiger increments (50-59%, 60-69%, 70-79%, 80-89%, and 90-99%). To test whether certain sized fragments were present more frequently than others, a oneway Chi-squared test was performed.

Interviews
To test whether there is a difference in the number of fishermen buying or collecting bait (Figure 2, Q3b), whether leftover bait is discarded more often than not (Figure 2, Q4), and whether bait is moved within and out of the estuary more often than not (Figure 2, Responses to the questionnaire were used in conjunction with observations of regeneration and data supplied by SANParks to explore if bait collection and regeneration can facilitate the persistence and anthropogenic dispersal of the species, as set out in Figure 2.

Interviews
To test whether there is a difference in the number of fishermen buying or collecting bait (Figure 2, Q3b), whether leftover bait is discarded more often than not (Figure 2, Q4), and whether bait is moved within and out of the estuary more often than not (Figure 2, Q3a), a one-way Chi-squared test was performed. All analyses were conducted in R Studio (version 4.2.1).
Data obtained from the survey together with the data from SANParks were used to estimate the potential scale of the problem. The number of bait collectors using Diopatra in Knysna Estuary was estimated using (2). This estimated value was then used to estimate the number of Diopatra caught per year using (3). The number of worms estimated to be extracted annually was used to determine the portion of Diopatra discarded per year using (4). The number collected per year and the reported frequency of breaking during collection was used to estimate the total number of potential breakages that can occur per year using (5). Lastly, (6) allows for the estimation of the portion of the worms that are large enough to regenerate and was based on the observations of in situ regeneration. n(bait collectors using Diopatra) = % prefering Diopatra × n(annual fishing effort from SANParks) (2) n(Dipatra caught per year) = n(mode number caught per person per day) × n(bait collectors using Diopatra) (3) n(discarded per year) = n(caught per year) × % discarded from interview (4) n(potential breakages) = n(caught per year) × frequency of breakage from interview (5) n(capable of regeneration) = n(discarded per year) × % that show regeneration from in situ observation (6)

In Situ Regeneration
From January 2021 to June 2022, a total of 594 specimens were collected, with 54.88% (n = 326) showing signs of regeneration. There was no difference in the incidence of regeneration between the two chosen sites. Of the 326 regenerating worms, 95.09% (n = 310) showed signs of anterior regeneration only ( Figures 3B and 4A,B), 1.23% (n = 4) showed signs of posterior regeneration only ( Figure 3C), and 3.68% (n = 12) showed signs of both anterior and posterior regeneration ( Figure 3A). The number of chaetigers regrowing in those regenerating anteriorly ranged between 7 and 17 (median = 10, mode = 9) (Figures 3B and 4A). All individuals regenerating anteriorly had 59-100% of the original branchiae intact. Regeneration was more prevalent in the individuals that had a higher percentage of original branchiate chaetigers intact (χ 2 = 886.04, n = 313, df = 5, p < 0.001) with most of the individuals falling in the 80-89% category ( Figure 4B). Furthermore, most individuals were regenerating eight to 13 chaetigers and had 80-89% of their original branchiae intact ( Figure 4B).

Fishermen Baiting Habit Survey
Seventy fishermen and bait collectors were interviewed throughout the Knysna Estuary. Of these, 35 were recreational and 35 were subsistence. Only 23 (32.86%) bait collectors and fishermen (16 recreational and 7 subsistence) selected D. aciculata as their preferred bait, and their responses were retained for further analysis. The respondents indicated that they collected a minimum of 1 and a maximum of 96 worms per day (median = 10, mode = 10).

Fishermen Baiting Habit Survey
Seventy fishermen and bait collectors were interviewed throughout the Knysna Estuary. Of these, 35 were recreational and 35 were subsistence. Only 23 (32.86%) bait collectors and fishermen (16 recreational and 7 subsistence) selected D. aciculata as their preferred bait, and their responses were retained for further analysis. The respondents indicated that they collected a minimum of 1 and a maximum of 96 worms per day (median = 10, mode = 10).
The fishermen who collected D. aciculata did not show a statistically significant trend towards moving bait from collection sites within Knysna (χ 2 = 3.522, n = 23, df = 1, p = 0.061), although a significantly greater number of fishermen do not move from bait collecting to fishing sits outside of the Knysna Estuary (χ 2 = 14.727, df = 1, n = 23, p = 0.0001) ( Figure 6A). Significantly more individuals collected their own bait compared to those who bought bait (χ 2 = 15.696, df = 1, n = 23, p < 0.0001) ( Figure 6B). A statistically significant proportion of the fishermen (n = 20) had bait leftover at the end of a fishing trip (χ 2 = 12.565, df = 1, p = 0.0004), and although more than half the fishermen interviewed indicated that they threw leftover bait back into the estuary or sea, this was not significantly more than the proportion who kept the worms to use on a later fishing trip or donated them to other fishermen (χ 2 = 4.9, df = 2, p = 0.0863) ( Figure 6C). Fishermen admitted to throwing away 10% to 50% of the bait they collected each trip (mode = 50% bait discarded per fisherman).
Fishermen noted that worms would break during bait collection significantly more frequently than not (χ 2 = 15.875, df = 4, p = 0.003) ( Figure 6D). Most fishermen indicated that worms broke 66-99% of the time during collection. The portion of the worm preferred as bait varied (χ 2 = 16, df = 3, p = 0.001) but most preferred the head or whole worms where possible. A statistically significant proportion of the fishermen (n = 20) had bait leftover at the end of a fishing trip (χ 2 = 12.565, df = 1, p = 0.0004), and although more than half the fishermen interviewed indicated that they threw leftover bait back into the estuary or sea, this was not significantly more than the proportion who kept the worms to use on a later fishing trip or donated them to other fishermen (χ 2 = 4.9, df = 2, p = 0.0863) ( Figure 6C). Fishermen admitted to throwing away 10% to 50% of the bait they collected each trip (mode = 50% bait discarded per fisherman).
Fishermen noted that worms would break during bait collection significantly more frequently than not (χ 2 = 15.875, df = 4, p = 0.003) ( Figure 6D). Most fishermen indicated that worms broke 66-99% of the time during collection. The portion of the worm preferred as bait varied (χ 2 = 16, df = 3, p = 0.001) but most preferred the head or whole worms where possible. A total of 19,954 bait collection efforts took place over a 12-month period [49] (Table 1). Using the proportion of bait collectors from our survey (32%) and SANParks (12%) ( (2) to (6)

Discussion
This study demonstrates that Diopatra aciculata has a great capacity for regeneration. Fifty-four percent of the individuals examined showed signs of anterior regeneration. However, the comparatively few chaetigers (less than 20% of the branchiate chaetigers) lost and being replaced anteriorly may reflect recovery from sublethal predation, rather than bait collection [36]. The natural predators of the genus include fish [29], birds [55], and crustaceans [56]. For example, the spotted grunter (Pomadasys commersonnii), a known predator of the species [29], is found in the Knysna Estuary. Additionally, the African sacred ibis (Threskiornis aethiopicus) was often observed feeding in the intertidal zones during low tide [46]. It is likely that these species, amongst others, are responsible for sublethal predation on D. aciculata in Knysna. On the other hand, only 3.05% of the population exhibited signs of posterior regeneration. A high incidence of posterior regeneration was linked to aggression among neighbouring worms in an aquaculture population when density increased to 2000 worms/m 2 [6]. This level of intra-specific competition is unlikely in the Knysna population where even at the estimated population numbers of 20-24 million individuals, density never exceeded 52 worms/m 2 (mean density 3.47 worms/m 2 ) [21].
Observations of numbers of branchiate chaetigers regenerating anteriorly, in conjunction with the original, intact, branchiate chaetigers, suggest that D. aciculata can have 20 to 70 branchiate chaetigers, depending on the size of the specimen, with the maximum nearly double what has been previously reported (20 to 40; [8]). Thus, if successful regeneration can only occur with at least half the original branchiate chaetigers intact [36], worms can survive the loss of approximately 10 to 35 of their branchiate chaetigers. Both anterior and posterior regeneration is greatly dependent on the presence of the branchiae [37]. The branchiae are extensions of the body wall containing loops of the vascular system that increases the surface area for gas exchange [57]. In tube dwellers such as Diopatra, the branchiae are located toward the anterior end where most water flows [57]. Therefore, regeneration would only occur in fragments that include the anterior portion of the worm that bears the majority of the branchiae. This is supported by our observations of anterior regeneration only being present when the amputation was before the 17th chaetiger, which is equivalent to at least 60% of the original branchiae intact (Figure 4). Similarly, during bidirectional and posterior regeneration, the smallest number of original branchiate chaetigers present were 21 and 39, respectively ( Figure 5). However, regardless of the total number of branchiate chaetigers present, no anterior regeneration was observed past the 17th chaetiger. This suggests that regeneration is only possible if amputation is before the 17th and after the 21st chaetiger. A similar trend was seen in D. neapolitana (15th and 25th chaetiger, respectively) suggesting that regeneration is limited by the specific chaetiger where amputation occurs, rather than the proportion of branchiate chaetigers lost [36].
Although bidirectional regeneration is possible for D. aciculata, neither D. neapolitana nor D. aciculata can incidentally reproduce asexually. In D. neapolitana, amputation in the mid-branchial region led to the death of both halves of the worm [35]. Similarly, no anterior regeneration was documented if more than 17 chaetigers were removed. We therefore conclude that as for D. neapolitana [36], it is unlikely that D. aciculata can withstand the damage inflicted by bait collection.
Only a small proportion (15%) of the regenerating worms collected displayed a level of damage that could be attributed to bait collection. This could be due to the methods used to remove bait, by hooking them out with piano wire inserted into individual tubes [21,30]. This targeted collection of Diopatra, may result in few individuals capable of regeneration being left behind inside the tube. Even when the whole worm is removed from the tube, it usually breaks into at least two pieces, leaving the anterior-most portion to burrow into the sand [46,47], if not picked up by the bait collector. Therefore, it is likely that worms showing signs of posterior or bidirectional regeneration had either escaped after removal or been discarded by fishermen.
If the ability to regenerate allows for the maintenance of the population despite bait collection, each individual harvested must leave behind a fragment capable of regeneration.
This would negate effects of bait collection, restoring population numbers, and ultimately creating an endless supply of bait ( Figure 2). Each fisherman can legally collect ten Diopatra worms per day [58] with subsistence fishermen active several times a week and recreational fishermen only on weekends and holidays [31]. Based on the results of the survey and in situ observations of regeneration, we estimated that less than 1% of the total population of D. aciculata is collected per year. Additionally, 20 of the 23 respondents had bait left over and discarded up to 50% of the bait collected, and it is therefore estimated that approximately 11,972 individuals are discarded per year. Of the discarded bait, 1765 fragments are big enough to settle and regenerate. However, this is a mere 7.37% of the total number that is collected annually, and unlikely to be enough to allow the population to not only withstand the pressures of bait collection, but to facilitate population expansion. On the whole, fishermen indicated that they complied with the daily allowable catch, but it is possible that they were underreporting their catches. For that reason, this estimated portion may be an underestimation. The small portion of individuals that lost more than 20% of their branchiae (15%) and showed signs of posterior and bidirectional regeneration (1.23% and 3.68%, respectively) further suggest that recovery after collection is unlikely. It is therefore clear that the sexual reproductive strategies of the worm must be robust enough to not only counteract the effects of predation and baiting, but also contribute to expansion. An investigation into the reproductive cycle and frequency of spawning is currently underway [59].
The staff of SANParks [50] report that recreational fishermen within a 100 km radius of Knysna frequently travel to Knysna to buy bait for use in other areas. This creates the risk of anthropogenic dispersal [43] if leftover bait that can survive and regenerate is discarded at the fishing site ( Figure 2). Only two fishermen admitted to buying and moving bait from Knysna, but this is probably an underestimation. Selling and by extension buying live worms is illegal [58], and interviewees may avoid incriminating themselves or guilty fishermen were not interviewed (see also [31]). Furthermore, the high unemployment rate in the area [60] is leading to an increase in subsistence fishermen and bait collectors illegally selling bait to recreational fishers. Therefore, this number is likely to increase in the future. Nevertheless, even if only a few worms are transported this way, the consistent movement of bait could create a high enough propagule pressure [61], resulting in the development of a self-sustaining population.

Conclusions
In conclusion, although Diopatra aciculata is capable of anterior and posterior regeneration, our data suggests that regeneration will not allow the species to withstand the effects of bait collection. However, the consistent movement of bait to other estuaries by fishermen can, under the right circumstances, lead to the development of new populations of the worm and the anthropogenically aided dispersal of the species. However, further research that investigates the dispersal capabilities of the species is required to strengthen this conclusion.