dear brendan especially

i am interested in first identifying all the cooch behar claves smaller than
that size
& in then carefully examining the nature & length of the perimeter of each
such candidate
in as much exact detail as is presently known or knowable
so as to further qualify or completely disqualify each of them

now i get the feeling there may only be a small handful of such qualifier
claves in all of cooch behar
& just knowing which ones they all are by size & name would be a big
advancement for me
for that would at least entirely surround the question for the first time

& then secondly
if i could somehow just pick your book
but preferably your brain too

Going by the book, (and you could always buy the book, Mike I have 2 copies left that need loving homes)
3476sq m comes to 0.859acres. Call it 0.86 acres.

Now there are a number of tiny enclaves for which I do not have separate areas, as they were considered fragments of a larger enclave.
There are 27 of these: India in Bangladesh: Banerjee's #48,49,50,44,45,46; Bangladesh in India: Banerjee's #2,4,5, 53,54,56,60,64,65,68,74,75,76,77,78,80,84,87,90,93,94
Of those for which separate areas are available:
Banerjee's # / Name / Area (acres)
India in Bangladesh:
52, a Fragment of Dahala Khagrabri enclave, 0.44
38, Natatoka, 0.26 [damn, i found a smaller one than 0.27!]
41, Bewladanga Chhat, 0.83
23, Najirgonja [Nazirganj], 0.77
15, Najirgonja [Nazirganj], 0.52
5, Garati, 0.79
98, Banskata, 0.77
110, Banskata, 0.28
105, Banskata, 0.64
61, Upanchowki Kuchlibari., 0.32
79, Panisala, 0.27

Bangladesh in India:
9, Nazirganj, 0.72
14, Upanchowki Bhajni 24, 0.71

Now the smallest perimeter to area is when a boundary is a circle. Thus the smallest possible area is 0.26ac = 1052sq m >>> circle of diameter 36.6m and hence a circumference of 115m
But none of the enclaves are circular. The largest ones have wiggly boundaries, but they all follow roads, rice paddy bunds and occasionally streams, though as stream and rivers change course a lot, what may originally have been riverbank boundaries, are now high and dry until diluviated again. It's a very complicated area hydraulically and cadastrally.
So we can approximate their perimeters to all intents and purposes by rectangles.
so 0.26ac = 1052m2 is a square of side 32.4m and so perimeter of 130m. Again a square is the minimum perimetered rectangle, so the boundary is likely to be slightly longer.
But as no decent maps are available, and the nature of Indian land surveying, land use etc does not allow for the same precision as in Baarle, we can't get down to cm length for the enclaves.
But 130-150m seems a reasonable guess.

As for the second boundary cross, I know where the information is, but it is unavailable to mere mortals such as we.

B