The definition given for imbricate with its origin in imbricare is interesting. Roman roofing tiles were of the "Under and Over" type ("Tegula and Imbrex"). A flat tray with curved up sides (the Tegula) was laid on the roof, and the joints between two trays were made rainproof by means of an inverted semi-cylindrical tile (the Imbrex) covering them. When the Romans ran off they took this secret with them and those few who still tiled, mostly monks, used flat tiles.

Flat tiles were laid with their side joints butted together, permitting water to fall through the joints. To stop this, any effective roof covering varied between 2 and 3 overlapping flat tiles thick to ensure that all the joints, lengthwise and lateral, were covered and the roof watertight. If you’ve ever laid roofing slate you will know what I mean.

Sometime in the 15th C the Dutch came up with the idea of linking tiles together laterally using an ogee or S-shape, rather than relying on their vertical overlaps to keep the water out. Because these tiles actually overlapped each other from side to side rather than butting, only 1 to 2 overlapping tiles were needed any point on the roof. These tiles became known in England as pantiles, probably from the Dutch word panne.

The word pantile was sometimes incorrectly used for flat Dutch paving tiles. A notable example is the Georgian spa Town of Royal Tunbridge Wells, some twenty miles from my home. The town has an historic area called ‘The Pantiles’ after the paving that was used there. Unfortunately the original paving was worn out long ago, but there are still samples in the local museum.

http://www.doot.co.uk/la24.htm

PS: I found that there are some other interesting shots in Mr Thompson's gallery.