This species with 2 subspecies parallels E decaryi with its 2 subspecies decaryi and spirosticha. It possesses leaves with a cylindrical cross section and a distinct center groove. The 2 subspecies are cylindrifolia and tuberifera. Subspecies cylindrifolia does not possess a center-caudex, but seed grown plants spread radially from a center growing point and it also spread through underground stems like E. decaryi. It can therefore forms a mat. Subspecies tuberifera froms a caudex if it is seed grown. The stems spread and branch out from the caudex but does not possess the underground stoloniferous habit. In time it can form a large caudex covered with a neat dense mat spread of linear leaves.
Above is the subspecies tuberifera with stems radiating from a center-caudex.
It is also a species from the dry Alluaudia-Didierea forest in southwestern Madagascar on limestone substrate. E. cylindrifolia v. cylindrifolia can easily be started from cutting and given time will grow true to form with stems radiating from a point. However, it is necessary to start E. cylindrifolia v. tuberifera from seeds to get a caudex plant. So far, i've not tried taking leaf cuttings of v. tuberifera to root but it should be possible. I did recall that one of my v. cylindrifolia stem cutting was rotten all the way to the base and only a rooted leaf was left. A new plant did eventually grow and forms a nice mat.