Tigers in Mysore


A tiger in Mysore (http://mysorezoo.info/)


The following opinion about our paper on tiger gene flow (ref here) was circulated in a Whatsapp group and was therefore, public. Parts of the text were questionable, but I limit my response to technical issues. The text says:

“… If you look at the resistance map closely, you will find that tiger can move through city of Mysore.

…they should have looked at other features…

And it is surprising to see from the data that the forest cover has also improved from open to dense forest…

The paper should have been read carefully. The legend in figure 2 is clear. There is a decrease in forest cover, not an increase.

Again, regarding “features”, the paper should have been read carefully. We took several variables (the same as other similar recent studies) and simplified to obtain the minimal adequate model. The minimal adequate model is different from other models and this may be due to the tiger genetic make-up, the genetic markers used, the landscape or everything together.

Now, figure 2 does not say that tigers can move through Mysore city. A gene flow resistance map is different from a resistance map extrapolated from movement. Without going into details (found in the paper, which again should have been read carefully), this is how it goes:

We found the same important variables (with correlation coefficients) as in other papers regarding gene flow (terrain, human disturbance, land use, etc), but some, such as land use were not significant. However frustrating it was, land use (vegetation, agriculture, settlements) was discarded. The lack of significance, again, may be due to sampling, markers, etc. But statistics caution that absence of evidence is not evidence of absence.

We were finally left with two significant variables (terrain and human disturbance). The effect of terrain on gene flow was different from that of Central India, but you have to remember that tigers are excellent dispersers and the effect of terrain is in any case, moderate. Then, tigers seem to avoid the “sight” of humans since the human disturbance variable has great effect at low intensity (a non-linear response with a small exponent to the variable).

Now we come to the interesting interpretation: are there tigers in Mysore? No. Except, maybe in the zoo. The figure 2 has its merits however. Viewed from the genetic make up of the genes sampled, and after systematic selection of variables, land use and cities have not been strongly “registered” by the genes. The genetic make up “does not see them”. This is surprising because radiotelemetry would have given entirely different results. But why is that so? Maybe natural selection has not allowed enough time to adjust to land use, since land use changes are recent when natural selection operates on a different time frame. Moreover, how can a species such as the tiger, being systematically killed outside reserves can evolve and adapt to human-dominated habitat? So the landscape seen “through the eyes” of the studied genes is similar to that of the ancestral line of the tiger. The extrapolation of results beyond the study area gives this very strange view of the landscape that only terrain and the sight of humans matter. This is not so outlandish and we do not say that tigers cross or should be reintroduced in Mysore.

This extrapolation was not the purpose of our paper. The paper discussed the difference between Central India and Southern India tiger gene flow in the view of maintaining population connectivity among reserves. The methods used there are solid and every year progress is made in the statistical model selection process. We may have learned something new here. And if it is wrong, it will be discarded by further studies.

Saying gene flow happens differently in different landscapes is rather rational and our paper, commented here by Ruth de Fries (who is an eminent landscape ecology specialist), was not interpreted otherwise.

Jean-Philippe Puyravaud