DISCUSSION


 At two areas at Ditlef Point (Fig. 2, 16) we collected a hun-
dred fifty Botany Plain, a Botany Narrow Handled, a Botany Un-
painted Plate, four Botany Griddle, fifty-six Hull Plain, one Hull
Painted Plate, a Bay Plain, a Coral Plain, and three Bordeaux
Plain sherds. The large percentage of Hull Series sherds in this
inventory suggests the major occupation occurred from very late
Magens I or very early Magens IIA intermittently to Magens IIB
times.
 Our collection from the site at Chocolate Hole (Fig. 2, 17)
contains thirty-four Botany Plain, one Magens Plain, one Bay Plain
and four Hull Plain sherds. The relatively small number of Hull
Plain sherds permits us to place this site in a generalized Magens
II period.
 On the southeast shore of Calvary Bay (Fig. 2, 18), we found
one Hull and two Botany Plain sherds. This site must have been
used for only a very short time.

 DISCUSSION

 Using the pottery type concept, I have analyzed the ceramic
collections made by us from both surface collecting and strati-
graphic testing on the islands of St. Thomas and St. John. As a
result two ceramic complexes have been delineated.
 The first or Coral complex represents, I believe, the first pottery
making and using culture of these islands. This complex included
thin, well-made and graceful vessels-sometimes white-on-red
painted, sometimes supplied with rim appendages or spouts
modeled in low relief-of the fine, sand-tempered, Coral Series plus
cassava griddles and utilitarian containers having wide handles
and made of coarse sand-tempered Hull and grit-tempered Magens
pastes. Simple rim points or lugs are rare but present. Red paint,
frequently limited to special areas such as everted lips, is fairly
common. This early ceramic complex existed during what we have
referred to as Magens I times.
 The second or Magens complex includes vessels of Botany, Bor-
 deaux, Magens, Bay and, rarely, Hull pastes. This complex existed
 for a long time during which styles in pottery decoration varied
 and became more complicated while careful attention to surfaces
 declined. I have tentatively delineated three phases in this develop-