The internal network of the termite mound is intricate, and understanding this architecture is important for understanding how the mound works as a gas exchange device.
Dissection has been the common way entomologists have tried to understand this structure. Dissection does not capture the organic integrity of the mound, however.
A better way is to borrow a technique from cell biology: recreate the 3D structure through a technique known as slice-and-scan. Section the structure into a series of 2D sections that can be reassembled into a 3D model.
This page outlines how Rupert Soar and I tried to do this for the mound of Macrotermes michaelseni.
Washing the matrix from a cast tunnel network
Filling a mound for slice and scan. A mound's internal aira spaces are filled with plaster (gray)
Endocasting simply means to fill the interior air spaces of the mound with a a durable material.
Two endocasting techniques can be used to reveal the structure of the internal tunnel network.
In the first, the matrix (in this case, soil) can be washed away to reveal the casts of the tunnel network beneath. This can be a quite delicate process, and can damage the fine structures of the tunnel network, but it can reveal a remarkable coherency of form for the large tunnel networks.
The second, called slice-and-scan, involves slicing away the structure, photographing the exposed section. Here, the mound's internal air spaces are filled, and the whole mound is encased in plaster to stabilize it. An XYZ gantry with a cutting wheel is then erected around the mound. The cutting wheel slices off the mound, and the exposed section face is then photographed, so that the tunnel stand out as white against the darker soil matrix. The cutting wheel is then lowered and the process repeated until the entire complex has been sectioned.
The serial photographs are then aligned in a computer, the outlines digitized and reassembled into a 3D virtual model.
Endocasting starts with a mound fill. We did this in 2008 with several mounds, with the generous on-site support of the staff at the Omatjenne Agricultural Research Station, near Otjiwarongo, where we work. Each mound was filled with several tons of a slow-setting plaster of Paris mix. Slow setting was required because the heat generated by such a large quantity of curing plaster could have distorted the structure. The plaster was was allowed to set and dry for six months.
From there, we washed the soil away from some of the mounds to reveal the tunnel network beneath. The results are impressive.
Slice-and-scan is a more challenging problem. The XYZ gantry was fabricated in England by Rupert Soar and John Webster, transported to Namibia and reassembled over a termite mound. The earth was excavated around it so that the nest could be sectioned all the way to the bottom, roughly 2 meters underground. The mound was then sliced at 1mm vertical intervals, with the cut face photographed each time.
Below are two time lapse videos showing the slice and scan machine in operation. The complete slicing of this single mound required roughy two months of effort, at least 18 hours per day and a crew of up to five people.
Below is a documentary video available on my YouTube channel (user: macrotermiteman) that describes the entire endocasting process, including an awesome movie of the reassembled slice-through of the mound.