ref:b1301-double tagging-gen-inf.doc

Double tagging, general information


Applying more than one tag or mark to the same individual is a common practice. An external mark or tag is frequently used to indicate the presence of an internal tag for example coded wire tag (CWT) in salmon where the adipose fin is cut off, or as in the case of DSTs (Data Storage Tags) when there can be an additional ID-tag for the indication of the presence of an electronic tag.


Double tagging experiments for estimates of tag retention.

Double tagging is also an important method for example to estimate tag loss. In such experiments two tags are attached to the same fish. These tags can be of the same type or different depending on the experiment. The assumption is made that the mortality rates, migrations etc. apply equally to all fish irrespective of the number of tags attached to the fish.

The probability of a tag having come off by time t, will be a function of time, say pt and

hence the probability of the tag not come off is (1- pt)


A) if the tags are the same

The probability of detachment should be the same

 2N = number of fish recaptured with two tags; Probability = (1-pt)2

0N = number of fish with no tags; Probability = pt2

1N = number of fish recaptured with one tag; Probability = 2 pt (1- pt)

 Then 1N / 2N = 2 pt / 1- pt

or 1N / (1N + 2N) = 2 pt / (1+ pt)



B) The tags are different

Then the probabilities of one or an other of two tags to become detached can be different.

 2N is the number recaptured with both tags; Probability = (1-A pt) *(1-B pt)

0N = number of fish with no tags; Probability = Apt * Bpt,

1NA is the number having lost tag of the type A; Probability = B pt* (1-A pt)

1NB is the number of fish with one tag of type B; Probability = Apt* (1-B pt)

Then 1NA/2N = Bpt/ (1- Bpt)


and 1NB /2N = Apt / (1- Apt)



Gulland (1963) redraws an example from Beverton and Holt (1957), showing the results of a double tagging experiment where fish was tagged with two Petersen tags one with silver wire and the other with stainless steel wire.

The results showed the probability of detachment (pt) for silver wire = 0.15 and for stainless steel = 0.025

In this case the Petersen tag with silver wire is 6 times more likely to be lost than the same tag if a stainless steel wire is used


A general model for tag-specific shedding rates is shown by Xiao, (1996) demonstrates a much more sophisticated model and it application to double tagging data from research on a carcharhinid shark.




Experiments have been done with electronic data storage tags(DSTs) where the DST is in the peritoneal cavity of the fish.

In this case the rate of loss of the DST was assumed to be very low compared to the rate of loss of a T-bar tag (floy tag).

If the rate of the loss of the DST is close to 0, the rate of loss of the T-bar tag is the ratio of recaptured fish with DSTs missing the T-bar tag (A) to the recaptured fish with DSTs where the T-bar is still in place (B).

With sufficient numbers of recaptured DSTs the ratio of A/B can be plotted against time from release of the tagged fish (Thorsteinsson unpublished results).




J. A. Gulland, 1963. On the Analysis of Double-tagging Experiments. Special publication ICNAF No. 3: 228-229.

J. A. Gulland, 1988. Fish Stock Assessment, FAO/Whiley Series on Food and Agriculture. Vol.1.

Beverton, R. J. H. and Holt, S.J. 1957. On the dynamics of exploited fish populations. Fish Invest., London.Ser. 2, 19.

Xiao, Y. 1996. A general model for estimating tag-specific shedding-rates and tag interactions from exact or pooled times at liberty for a double tagging experiment. Can. J. Fish. Aquat. Sci. 53: 1852-1861 (1960).

Comments to:
All rights reserved © Marine Research Institute

Last updated: