Biogeographic transitions for tropical Butterflyfish around South Africa

Just to give an example of what I have in mind, I took the distribution range of the 14 species of tropical Butterflyfish (most of them from the genus Chaetodon) around South Africa. I excluded the one endemic C. marleyi as its distribution range (Maputo to Lambert's Bay) suggest that it is not tropical. Then, I recorded the presence (1) / absence (0) of each species in 20 locations along the east and south coasts. I then performed a logistic regression pooling the data from all the species. I obtained the graph below that gives the probability of occurrence of tropical Butterflyfish. The dots represents the presence/absence data in different locations and the red curve represents the regression model.

So, as a preparation for the next discussion group, we can try and interpret the graph. As guidelines, try and answer the question below:
(1) Considering the trend of the regression model, can you suggest the location I used for the "origin" (coordinate 0 on the x axis)? and the last location (x ~ 3500 km)
(2) Considering the position of T1, T2 and T3 on the graph, what do they represent in term of the occurrence of the Butterflyfish?
(3) What are the biogeographic interpretations one can make in term of species distribution using these three lines?
Try answer these questions and if you have any questions, please contact me by email or write a comment.
Enjoy!