Gastonia was a medium-sized ankylosaur. In 1998, Kirkland estimated its length at . The skeleton mount has a length of with a hip height of . It is difficult to determine distinguishing traits of
Gastonia because its affinities are uncertain. However, in 1998 Kirkland established three characters which were unique for the Ankylosauria as a whole and thus likely
autapomorphies, unique derived traits. On the midline of the front snout, the bony core of the upper horny beak, a broad, gradually curved, notch is present between the snout bones, the praemaxillae. The bony nostrils are placed far to the rear. At the underside of the braincase, the basisphenoid, the basipterygoid processes are longitudinally stretched.
Gastonia had a flat and, even for an ankylosaurian, very broad rump, the belly strongly protruding between the short powerful limbs. The tail was moderately long and lacked a tail club. The neck was relatively long and the skull probably rather small. The skull is somewhat elongated and pointy, measuring 295 by 283 millimetres in the holotype. The top profile of the skull is convex, the rear skull roof curving below the level of the upper rim of the eye sockets. The quadrate is very strongly inclined to the rear. The occipital condyle, the contact with the neck, is obliquely pointing to below, an ankylosaurid trait causing the head to be pointing downwards. The beak is toothless. The tooth rows of the
maxillae are rather straight and each consist of fifteen to sixteen small teeth, lacking a true
cingulum, swollen basis. There is no armour on the snout. More to the rear the bony tiles of the skull roof, the
caputegulae, are rather indistinctively patterned though a small central plate on the
parietal bones is visible. The squamosal horns, at the rear skull corners, and jugal horns, at the cheeks, are small. The
sacrum consists of three sacral vertebrae, with a caudo-sacral vertebra behind them. Stiffened rear tail vertebrae or a tail club are absent. The shoulder blade resembles that of nodosaurids in possessing an
acromion that is blade-like, but differs in the acromion originating from the front edge, not the outer side of the scapula, not being wrapped to behind and not ending in a knob. It is thus not a typical nodosaurid "pseudoacromion". The
coracoid is square in form, a typical ankylosaurid trait. The humerus has a large deltopectoral crest extending downwards to the middle of the shaft, a derived trait. The
ulna is very robust with an enormous
olecranon. The shinbone is very expanded at both the upper and lower end.
Gastonia was protected by
osteoderms, skin ossifications. The neck was covered by at least two bone rings. Usually in ankylosaurs these have the form of "halfrings" leaving the underside unprotected, but with
Gastonia only two segments seem present, one at each side of the midline, causing Kirkland to refer to them as "quarter rings". Each segment had a pointed keel and a hollow underside. Kirkland stressed that the rump armour was hard to reconstruct because it had not been found in articulation. The sides of the thorax seem to have been covered by about five pairs of large flat triangular spikes. They are recurved and have a deep groove in the rear side. They gradually decline in length to behind, the groove becoming relatively shorter and the base length increasing. According to Kirkland the function of the groove was to receive the front edge of the next spike. Other large flat spikes found, lacked the groove. They were often very curved, the point at a right angle with the base. Kirkland assumed these formed two vertical rows, one at each side of the rump midline. Lower triangular spikes he placed at the sides of the tail, again gradually decreasing to the rear. In between the horizontal and vertical spikes of the rump probably rows of osteoderms were present having the profile of a droplet, with a vertical point at the broader end. The top of the tail had oval plates. The hip region was covered by a large pelvic shield consisting of fused osteoderms. These were patterned as
rosettes with a larger plate in the middle, surrounded by at least two rings of smaller plates. Kirkland assumed that four pairs of triangular spikes covered the sides of the pelvic shield also, but this was denied by Paul. The area between all these larger elements was covered by small ossicles, round bony scutes with a diameter of up to two centimetres, hundreds of which have been discovered.
G. lorriemcwhinneyae differs from
G. burgei in having a flat skull roof, shorter and narrower paroccipital processes, a postacetabular process that is only 36% the length of the preacetabular process, and an ischium that has an unkinked, smooth bottom edge. ==Phylogeny==