Manchester Memoirs, Vol. xlvii. (1903), No. 4. u 



galleries with inter-connections on a fixed and universal 

 plan. It is true that rarely the spiral tunnels might give 

 one this idea if onl}^ one fortress were dissected, but after 

 dissecting some 300 of them I can only say that no two 

 were exactly alike. I give a figure {Fig. 1 1) of an extremely 

 rare case of the spiral tunnel assuming the form of an 

 "upper circular gallery." This fortress was a very large one, 

 and I much regret that time would not allow my investi- 

 gating the very numerous exits from the complete lower 

 circular gallery into the meadow. It is noticeable that 

 the upper runs do not communicate with the nest or with 

 the three tunnels leading from the nest. 



Now, as has been described, it often happens that from 

 runs at the foot of the fortress several up-shafts will be 

 found. These often become connected at their base in 

 the following manner : — The frequent tunnelling close to 

 the fortress often loosens a large portion of turf (often 

 18x12 inches in area), and as this is heaved up a 

 connection is opened from run to run, and becomes what 

 the books call the " lower circular gallery." This is very 

 seldom complete, as in Fig. 11. If the turf is very loose, as 

 we find it on peaty or marshy land, much larger pieces 

 are heaved up {Fig. 14), and I once found a nest of young 

 beneath the unbroken turf which was slightly raised 

 [Fig. 13). The average fortress is one foot in height and 



Fig. 13. — N. — Nest under turf. 



