An interesting statistic in any tagging study is the tag loss rate. It is a vital element where the study looks to quantify mortality, movement, recruitment or abundance through tagging rather than ageing. While not a mission-critical consideration for me, a high loss rate can reduce long-term recaptures and possibly bias some growth analyses (see references below). There is also a logical connection between growth rates and factors that lead to tag loss. So while tag loss estimation was not absolutely vital for my work, I have some useful data and resulting observations.
In my previous two blog posts I reported tag recapture rates of around 8%-20% for Lunella torquata and 12%-50% for Turbo militaris. These are good in comparison to many fisheries population dynamics experiments but not particularly remarkable for relatively sedentary demersal shellfish gathered by hand in shallow water. Those figures would have been higher if there was no overgrowth on shells rendering tags unrecognizable, or if snails did not lose their tags.
I have not examined subsets of tag loss data for individual sites or seasons, nor have I compared males and females but the information is available to anyone interested in it.
Because I double-tagged and triple-tagged most snails, and I had many recaptures, I can estimate a simple tag loss rate and the overall results are shown in Table 1 below.
% of double-tagged snails recaptured with 1 tag (1 tag missing) | % of triple-tagged snails recaptured with 2 tags (1 tag missing) | % of triple-tagged snails recaptured with 1 tag (2 tags missing) | |
Lunella torquata | 21.1% (641) | 7.0% (442) | 5.4% |
Turbo militaris | 25.8% (357) | 28.5% (722) | 6.8% |
Table 1. Overall tag loss rates for double and triple-tagged snails using combined data for all sites and all times at liberty. Figures in brackets are the numbers of snails double and triple-tagged respectively.
Based on these overall data, for triple-tagged Turbo militaris, 28.5% of 28.5% equals 7.84% which is not far from 6.8%, so it would not seem that losing a second tag is correlated with losing one tag. I assume them to be independent events. This seems not to be the case for Lunella torquata, where 5.4% is much greater than 7% x 7%. I will discuss this further below because tag loss is a time-variable event, so these simple indicators are not conclusive.
Temporal trends are shown in Figures 1 and 2 below.
In these two figures, the values are percentages – actual numbers of recaptures are smaller for longer time periods so some percentages may not be reliable.
Figure 1. Time-variable tag loss rate for Lunella torquata. Combined
recapture data from all sites showing months between tagging and recapture.
Figure 2. Time-variable tag loss rate for Turbo militaris. Combined recapture data
from all sites showing months at liberty.
Comparing the two figures, it is noteworthy that the single tag loss rate between 5 and 10 months is very similar for both species at around 25% to 30%, and that for both species, the percentage of recaptures that lost 2 of 3 tags is also similar at below 10%. As for the overall rates above, probabilities are multiplicative, so since 25% of 25% is 6.25%, I conclude that losing one tag and losing a second tag are independent events and not correlated.
Tag loss increased to 30%-40% after 12 months, so it is understandable that it might have contributed significantly (along with fouling, mortality and migration) to my low numbers of recaptures for both species after about 15 months.
As outlined in my blog post on methods, tags were cut from various coloured plastics and attached with cyanoacrylate glue (“SuperGlue”). The plastics probably had different chemical compositions so there may have been variability in attachment success. Also, there are probably different formulations in the various brands of glue. Some that I used were more gel-like than others. The gel glues were better for filling the tiny corrugations under the tags but took longer to set than the thinner glues, which ran under the tags more easily by capillary action.
Differences in tags and glue might explain some variability, but the similarity and consistency in the figures above suggest that the nature of the tag, the quality of the glue, and the integrity of the tagging process itself are probably not confounding factors. Ideally, if tag loss was critical to the experimental design for the research question being investigated, all tags would be of the same plastic and the glue would all be from a single manufacturer. Kate Seinor used numbered tags that were designed to tag queen bees in her work in northern NSW on Turbo militaris, and these would be preferable in some respects, especially for tracking multiple recaptures. I took the cheap approach since my work was self-funded, and I could cut the tags to very small sizes to fit shell sculpturing. This was less of an issue with the smoother-shelled Turbo militaris for which I expected higher tag loss but that was not evident in the data.
It is worth considering other reasons for tag loss. I suspect that the main reason is abrasion since these snails crawl around in crevices and under boulders. Turbo militaris in Mahon Pool, where I could observe them easily at night, come out of their undercut crevices and graze across the bottom of the pool and then return to their crevices before dawn. (This may not apply to the very small juveniles, which might explain the non-linear growth rate shown in my previous post.)
These movements might explain the slightly increasing trend over time. The almost immediate high rate in less than 2 months might be due to post-release movement. Except for Mahon Pool, I did not return snails to the exact crevice, rock or hole they were collected from – only the general 500m2 area. Many undoubtedly searched quickly for a more favourable spot than where I placed them. This might also explain the very slow increase over time. I suspect many do not move much, and only some go exploring over time. This was found for abalone in Japan where many stayed in their “home stuck places”.
If there is post-release behavioural distress of some kind, (for example vulnerability to predators or unacceptable turbulence) it could affect feeding and thus growth. There are clear disturbance rings at the length of tagging in many of my snails, which would suggest that any growth rates I publish may be under-estimates. Disturbance rings have been found in Turbinid gastropods (e.g.,Sire and Bonnet,1984), abalone (McShane & Smith,1992) and indeed other molluscs such as squid (Jackson,1990). They are quite common (Hollyman et al,2018).
I have not documented disturbance rings, which can be due to tagging, storms or other episodic events. I consider the likely impacts on my findings for my research questions are:
- What is the growth rate at various sizes? Possible size-related impact.
- Does it vary seasonally? Possible impact.
- Are sites different? Unlikely
- Does growth vary between years? Very unlikely.
- Is growth different for the 2 species? Very likely
- Are males and females different? Possible but unlikely.
As with all ecological research, the process is not linear but circular. Results always lead to further questions (see Underwood, 1997). All my recaptured shells (with operculums) have been handed over to NSW Fisheries (Department of Primary Industries) and are available for study by any prospective workers who might wish to spend more time looking through a microscope eyepiece than me.
Using a single-tag loss rate of say 25%, losing both tags from double-tagged snails or all tags from triple-tagged snails would contribute only about 6.25% or 1.6% respectively my tagged snails not recaptured. For Lunella torquata this was 89%. For Turbo militaris at ocean reefs this was 82% and for Turbo at Mahon Pool it was 50%. This means non-recovery was due more to mortality, migration and non-recognition (growth of epibionts on shells) and not so much due to tag loss.
Note that Mahon Pool is an anomaly for Turbo militaris juveniles. Densities are much higher than I found on open reefs. The high recapture rate, low tag loss rate and relatively low growth rate for juveniles at this site suggest that they may be more sedentary in the back of crevices than larger snails and may not be moving out to graze as much. I did not find many large snails in the pool and I think they actively move out of the pool at about 60mm to 80mm.
REFERENCES
Hollyman P R, V V Laptikhovsky and C A Richardson. 2018. Techniques for Estimating the Age and Growth of Molluscs: Gastropoda. Journal of Shellfish Research 37 (4), 773-782. https://doi.org/10.2983/035.037.0408
Jackson G D, 1990. Age and Growth of the Tropical Nearshore Loliginid Squid Sepioteuthis lessoniana Determined from Statolith Growth-Ring Analysis. Fishery Bulletin 88(I):113-118
McShane P and Smith, 1992. Shell Growth Checks Are Unreliable Indicators of Age of the Abalone Haliotis rubra (Molluscs : Gastropoda) Aust. J. Mar. Freshwater Res. 43, 1215-19
Seinor K, Purcell S W, Malcolm H, Smith S D A and Benkendorff K (2023). Extended and spatially asynchronous reproductive periodicity in a harvested, warm-temperate rocky-reef gastropod (Turbinidae). Fisheries Oceanography 1-12 https://doi.org/10.1111/fog.12653
Sire J Y and P Bonnet. 1984 Croissance et structure de l’opercule calcific du gastropode polynesien Turbo setosus (Prosobranchia: Turbinidae): determination de l’ age individuel (Growth and structure of the calcined operculum of the Polynesian gastropod Turbo setosus (Prosobranchia: Turbinidae): determination of individual age) Marine Biology 79:75-87
Underwood, A.J., 1997. Experiments in ecology: their logical design and interpretation using analysis of variance. Cambridge university press.
Xiao 1994. Growth Models with Corrections for the Retardative Effects of Tagging. Canadian Journal of Fisheries and Aquatic Sciences. February 1994. https://doi.org/10.1139/f94-027