Samples of rabbit burrows were excavated and measured at four areas around Edinburgh. Of these, 31 burrows over 2 m long were analysed for internal structural relationships and the effect of soil composition and slope of the ground on their form. Seven variables were measured for each burrow—total length, number of entrance holes, number of junctions, number of ends, average length of sections between holes, junctions and ends, average depth and maximum depth. There are three main trends in burrow variation—size, an inverse relationship between relative number of holes and average depth, and an inverse relationship between average section length and relative number of junctions. Burrows in sand have relatively fewer holes and junctions, longer element lengths, a higher average depth and three times the enclosed volume per hole. Burrows dug into slopes have a higher average depth. Areas differed significantly in the relative number of holes and junctions, and average depth. It was concluded that an apparently complex structure could be summarized in terms of relatively few components. The chief of these, size, was largely independent of the soil and site characteristics, whereas the remaining two were dependent on the soil in which the burrow was dug. This may have implications for the ecology and behaviour of rabbits in different areas, and is relevant to the success of some rabbit control procedures such as burrow fumigation and warren ripping.