seaturtle.org : MTN : ARCHIVES : Sign In

Marine Turtle Newsletter 22:11-15, © 1982

What Double Tagging Studies Can Tell Us

N. Mrosovsky and Sara, J. Shettleworth
Departments of Zoology and Psychology, University of Toronto, Toronto, Canada M5S 1A1

Double tagging, as advocated in the first Marine Turtle Newsletter (1976), is not merely a way of obtaining more returns. It also enables estimates to be made of tag loss. This knowledge may be helpful in tackling other questions such as what are the chances of turtles re-migrating in future years to their nesting beaches. What is it that is missing in the "lost majority" (Carr, 1980) of turtles tagged at Tortuguero? Is it the turtles or the tags?

Estimating tag loss. First we consider the case of dissimilar tags, or tags which are the same but may have a different probability of loss (e.g., one on the back and one on the front flipper, or one tag and a notch on a acute). All estimates apply only to a time span. Also, if concerned with tag loss itself, one must assume that tags have an equal probability of being noticed. When a trained person systematically inspects a turtle on the beach, this may be a safe assumption. When a fisherman who is unaware of the tagging programme pulls a turtle out of his net, visibility of a tag may be important (Cornelius and Robinson, 1982). Where returns are predominantly from fishermen, it may be preferable to speak of tag irrecoverability rather than tag loss. This would encompass tag loss, tag visibility and the chances that a tag will be returned. Irrecoverability could be calculated in the same general way as given here for tag loss:

Let P_a be the probability of tag type-a being lost
Then 1-P is the probability of tag type-a remaining on
Let P_b be the probability of tag type-b being lost
Then 1-P_b is the probability of tag type-b remaining on.

Thus, for example, the probability of tag type-a remaining on while tag type-b is lost is P_b (1-P_a).

Let T be the total number of turtles initially double-tagged
Let R (unknown) be the proportion of T returning
Let N₀ be the number of returning turtles that have lost both tags
Let N_1a be the number of returning turtles with just tag type-a on
Let N_1b be the number of returning turtles with just tag type-b on
Let N₂ be the number of returning turtles with both tags on.

From equations 3 and 4:

Rearranging:

Similarly

If the tags are the same and are assumed to have an equal probability of loss, then the expression may be simplified. If N₁ is the number of turtles with one tag, then N₁/2 may be substituted for either N_1a or N_1b in equations 5 or 6. And P (the probability of losing one tag) may be substituted for either P_a or P_b in the same equations:

Rearranging:

Examples of tag loss/irrecoverability. 1) Green (1979) gives data from a double tagging study on a population of green turtles resident around the Galapagos Islands. These were tagged with monel metal on the front flipper and plastic (Rototag) on the hind flipper. After 4 years, 63 turtles had been seen again with both tags, 45 with just the plastic tag and 4 with just the metal tag. From equations 5 and 6:

These probabilities are for a time span of 101-1,000 days from the time of tagging. Fortunately Green also gives the time between tagging and recapture. Although the data are few for the longer intervals, they can be used to illustrate how probabilities of tag loss only refer to particular time spans:

2) Cornelius and Robinson (1981, 1982) double tagged 2415 olive ridleys on the west coast of Costa Rica. A metal tag was attached to one of the front right flippers and a plastic tag (Allflex) to any of the four limbs: 97 turtles were seen again with both tags on, 117 with just the plastic and 150 with just the metal. From equations 5 and 6:

Comments on examples: Neither of these studies demonstrate the superiority of one type of tag material. It may have been that the hind-foot location rather than the plastic was the important factor in the Galapagos study. With leatherbacks, as first noted in the Marine Turtle Newsletter (Hughes, 1978, see also in press), switching tag location from the front to rear flipper was followed by a jump in recovery rates. With the double tagging of the olive ridleys in Costa Rica, the time span of the study needs to be made explicit and considered. Perhaps over several years one of the tags used will have superior staying power. Also both the location and the colour of the plastic tags varied. Yellow tags had very poor recovery rates (Cornelius and Robinson, 1982). Before drawing definite conclusions from this study it will be necessary to wait for further details on these variables and their interrelationships. What is asserted here is that expressing results in terms of the probability of tag loss is an easy, useful and standard way of describing and comparing tag loss as a function of material, colour, location and time after application. Also the surest progress will be made by studies that do not confound variables. Whether a tag is put on the left or right flipper might even make a difference. Leatherback turtles in French Guiana more often have injured left than right flippers (Fretey, 1981). All variables except the one under consideration should be equated.

But one substantive point is evident. In two separate studies, on different species in different parts of the world, there was a probability of about .5 that a monel metal tag on the front flipper would be lost. This is still the most commonly used tagging method in turtle research. Clearly if around 50% of the tags fall off, it is hardly satisfactory, as had been recognized qualitatively (Balazs, 1982).

Scar method: At Tortuguero tag loss of monel metal tags has been estimated at 26.4% from scars left by missing tags (Carr, 1980). Is the tagging method superior at Tortuguero, or do green turtles from that population live in an environment conducive to tag retention or are scars unreliable as a way of assessing tag loss? Many people have wondered if scars may heal over altogether. Validation of the tag scar method is needed. This could easily be done by double tagging. Any turtles found with no tags should either have two scars or none. The same general formulae as given above could be used to assess the chances of scars becoming unrecognizable. If single tagging had been done in the area previously, then one could look for scars on those turtles from the double tagging experiment that returned with only one tag. Richardson et al. (1978) have been both double tagging loggerheads for many years and recording tag scars. They may well have relevant data.

Remigration Rates

By definition: TR = N_o + N_la + N_lb + N₂

The number of turtles with no tags is composed of those that have lost both tags (N₀) and of neophytes. N₀ cannot be measured directly. Therefore for N₀ we substitute PAPJR from equation (1).

TR = P_a P_b TR + N_la + N_lb + N₂

Rearranging and cancelling:

In the case where the tags are the same and are assumed to have an equal probability of loss, calculations may be simplified by substituting P (the probability of losing one tag) for both P_a and P_b and by substituting N_1a (the number of turtles with one tag) for N_1a + N_1b in equation 8:

These formulae look attractive but their application for calculating remigration rates is debatable. One problem is that turtles come back to the nesting beach after varying numbers of years. One cannot lump tag returns for a number of years (to ensure that each turtle has bad a chance of coming back and being seen in one year or another) and still use equation 8 because the probability of tag loss is increasing over those years. P_a and P_b in equation 6 are not constant. Instead it would be necessary to treat each year separately, calculating both probability of tag loss and remigration rate that year. Remigration rates for a number of separate years could then be summed. For this approach it might be simplest to tag intensively over a relatively short time span, but if not enough turtles could be tagged in one season, then the number of years since tagging could be used for assembling remigrant turtles into different groups for assessing tag loss and remigration rates.

Presumably calculations of the kind given in this note are to be found elsewhere. No claim to originality is made. We hope, however, with double tagging becoming widespread, that it may be useful for turtle researchers to have formulae for deriving tag loss readily available and that they may be stimulated to explore further how data from double tagging studies can be made instructive (see Seber, 1973 and Eberhardt et al., 1979 for further reading).

Balazs, G. H. 1982. Factors affecting the retention of metal tags on sea turtles. Marine Turtle Newsletter 20:11-14.

Carr, A. 1980. Some problems of sea turtle ecology. Amer. Zool. 20:489-498. Cornelius, S. E. & Robinson, D. C. 1981. Pacific coast of Costa Rica: tagging programme. Marine Turtle Newsletter 18:13.

Cornelius, S.E. & Robinson, D. C. 1982. Abundance, distribution and movements of olive ridley sea turtles in Costa Rica, II. Report, U.S. Fish & Wildlife Service, Albuquerque, New Mexico, pp. 1-40 plus figs. and tables.

Eberhardt., L. L., Chapman, D. G. and Gilbert, J. R. 1979. A review of marine mammal census methods. Wildlife Monogr. 63, pp. 8-11.

Green, D. 1979. Double tagging of green turtles in the Galapagos Islands. Marine Turtle Newsletter 13:4-9.

Hughes, G. 1978. Tagging leatherbacks. Marine Turtle Newsletter 6:4.

Hughes, G. R. in press. Nesting cycles in sea turtles-typical or atypical? In: Biology & Conservation of Sea Turtles (Bjorndal, K. A. editor), Smithsonian Institution Press, Washington D.C.

Mrosovsky, N. 1976. The tag loss problem. Marine Turtle Newsletter 1:3-4.

Richardson, J. I., Richardson, T. H. & Dix, M. W. 1978. Population estimates for nesting female loggerhead sea turtles (Caretta caretta) in the St. Andrew Sound area of southeastern Georgia, U.S.A. Florida Mar. Res. Publications 33:34-38 (Florida Department of Natural Resources).

Seber, G.A.P. 1973. The estimation of animal abundance and related parameters. Griffin: London, pp. 94-96.

Fretey, 7. 1981. Note Sur les traumas observes chez des tortues luths adultes Dermochelys coriacea (Vandelli) (Testudines, Dermochelyidae). Rev. fr. Aquariol. 8:119-128.