*** START OF THE PROJECT GUTENBERG EBOOK 69783 ***

                  *       *       *       *       *

                A Synopsis of the Palms of Puerto Rico.


                             BY O. F. COOK.


  [Reprinted from the BULLETIN OF THE TORREY BOTANICAL CLUB, 28. Oct.,
                                 1901.]

                  *       *       *       *       *




                 A Synopsis of the Palms of Puerto Rico

                             BY O. F. COOK

                          (WITH PLATES 43–48)


The following systematic notes have been accumulated in connection with
economic studies of Puerto Rico[1] palms, and although the list is
doubtless still incomplete, the printing of it may be justified as a
means of securing at least provisional names needed for reference
purposes in connection with other publications of a non-systematic
character.

The palms may well be considered a very refractory group when handled by
the conventional methods of systematic botany. Difficult at once to
collect or to study from dried material, they are commonly neglected
both in the field and in the herbarium, with the result that literature
is scanty and unsatisfactory. A very large proportion of the
descriptions are entirely inadequate for the identification of species,
and there has been much lawlessness and diversity in the application of
generic names, as will appear from some of the instances discussed
below. Difficulties of description and classification have also been
multiplied by the fact that the palms are such peculiar plants that
analogies and criteria borrowed from other families are often
inapplicable and misleading. Moreover, the terminology of parts and
characters has not been developed to the point where the expression of
observed differences is easy, and available language often fails
completely to suggest the significance of the characters used. Thus the
fibers into which parts of the leaf-bases of many palms are resolved
afford many diagnostic characters, for which we have no parallels in
other groups of plants.

A compensating advantage may be drawn, however, from the definite and
often very limited geographical distribution of the species of palms.
Thus, although Puerto Rico is a relatively small island, several of the
indigenous palms have apparently ranged in nature over but a small part
of it, and a locality definitely indicated would often go further toward
establishing the identity of a species than much of the descriptive
matter prepared for this purpose. For the present, at least, the
geographical idea should be kept uppermost in systematic studies of the
palms, since it is generally much easier and far more logical to extend
the limits of supposed distribution and unite supposed species, than to
cope with the confusion caused by the miscellaneous reporting of species
far outside their natural ranges.

From the popular standpoint another serious inconvenience of the
systematic literature of palms arises from the fact that it is based so
largely on floral characters that even the botanical traveler might need
to wait months for the blossoms and then climb the trees or cut them
down before being able to secure a clue to botanical names or
relationships. But however necessary refinements of formal characters
may be in presenting classifications or monographs of large groups, more
obvious differences may still be adequate for distinguishing between the
species, genera and families represented in a limited flora like that of
Puerto Rico. In the present paper use is made therefore of obvious
external differences, not only because of the greater convenience and
utility of such facts in field study but also in the belief that with
the palms, at least, the vegetative, habitat and ecological features are
often quite as important for diagnostic purposes as the more technical
and conventionalized characters to which botanists are accustomed in
dealing with other natural orders.

As will be apparent from some of the following systematic notes, the
generic nomenclature of the palms is in a condition closely comparable
to that now known to obtain among the myxomycetes, fungi, hepaticae and
ferns. Possibly the palms have suffered more from neglect and
carelessness than other groups of flowering plants, but it can no longer
be maintained that the practical defects of former taxonomic methods do
not exist in the phanerogams as well as in the cryptogams, and it
becomes obvious that the enactment of different nomenclatorial
legislation for these two subdivisions of the vegetable kingdom would be
unreasonable and inconsistent.

The present list records twenty palms from Puerto Rico, of which three
are introduced and seventeen are supposed to be native species. As may
also be inferred from many other groups of plants Puerto Rico appears to
be a rather remote corner of the Antillean region, which many types
present in Cuba and Jamaica did not reach, whether by reason of greater
distance from the continent or because of an earlier interruption of
land communication. The native palms of Puerto Rico may thus be said to
represent a distinctly Antillean or Caribbean series, only _Acrocomia_
and _Bactris_ being known to have a wider distribution.

The list of introduced palms, consisting of the date, the cocoanut, and
the betel, might have been somewhat increased by canvassing ornamental
gardens, but it does not appear that any other introduced species has
been put to any useful purpose or has escaped into general culture,
certainly a remarkable fact when we consider the number and importance
of the economic palms of other tropical countries.

Finally, it may be well to note here that several palms have been
reported from Puerto Rico which probably do not exist in the island; at
least their occurrence is not supported by adequate evidence. Thus Mr.
R. T. Hill, of the United States Geological Survey, mentions (Bull. U.
S. Dept. Agric., Division of Forestry, 25: 1899) as occurring in Puerto
Rico seven palms, as follows: _Cocos Mauritia_, _Oreodoxa oleracea_,
_Cocos nucifera_, _Martinezia caryotaefolia_, _Mauritia flexuosa_,
_Oreodoxa regia_, and _Caryota_ sp., of which list only _Cocos nucifera_
and _Oreodoxa regia_ appear to have been justified.

The reference to _Oreodoxa oleracea_ is supported by the botanical
authority of Professor Drude, but the specimens identified by him as
_Oreodoxa oleracea_ (Sintenis collection, no. 1525) and sent from the
Berlin Botanical Garden to the National Herbarium and to the New York
Botanical Garden are not _Oreodoxa oleracea_, but belong to the new
genus _Acrista_ described below, while a specimen collected by Sintenis
(no. 5749) at Aguadilla and sent out from Berlin as an _Attalea_ or
related genus is not even a cocoid palm but _Areca catechu_, the betel
nut of the Malay region.

The existence of numerous tubercles on the roots of a young specimen of
the royal palm of Puerto Rico is a fact of biological interest and
possible economic importance. It was, however, noted so nearly at the
end of our last visit that further studies were not practicable, but
barring possible nematodes or other pathological causes for the
tubercles it appears that we must add palms to the Leguminosae,
_Podocarpus_, _Alnus_, and _Cycas_ as plants which have, as it were,
domesticated nitrogen-collecting soil organisms.

The field notes, specimens and a considerable series of illustrations
for publications of the Department of Agriculture were secured during
two visits to Puerto Rico, the first in November and December, 1899, the
second in June and July, 1901. The photographs are the work of Mr. G. N.
Collins.


                          Key to the Families

  Leaves fan-shaped; branches of inflorescence subtended by spathes.

                                              Family SABALACEAE, p. 529.

  Leaves feather-shaped; spathes few, not subtending the branches of the
      inflorescence.

        Leaf-divisions v-shaped in section, concave above; trunk rough
            with leaf-bases or prominent diamond shaped scars.

                                            Family PHOENICACEAE, p. 528.

        Leaf-divisions inverted v-shaped in section, convex above; trunk
            smooth or the leaf-scars ring-like and not prominent.

        Leaf-bases long-sheathing, green and fleshy, finally split down
            the side opposite the midrib permitting the leaf to fall;
            fruits with fleshy, fibrous or woody endocarps.

                                               Family ARECACEAE, p. 546.

        Leaf-bases sheathing only while young, with maturity separating,
            except at the midrib, into a dry fibrous network which must
            tear or decay before the leaves fall; fruits with a stony
            endocarp perforated by three foramina.

                                                Family COCACEAE, p. 558.


                          Family PHOENICACEAE

This family contains a single genus of old-world palms usually
associated with the fan-leaved series, and differing from all other
feather-palms by having the concave side of the leaf segments turned
upward.


              PHOENIX DACTYLIFERA Linn. Sp. Pl. 1188. 1753

The date palm was probably introduced into Puerto Rico in the early part
of the Spanish occupation of the island, and isolated trees are to be
found in many localities especially in the vicinity of the larger towns.
The climate is, however, too cool and too moist to permit the fruit to
ripen properly, and there is apparently no inducement for planting in
large quantities.


                           Family SABALACEAE

Although forming no conspicuous part of the palm vegetation of the
island the fan-leaved species seem to be more numerous than those of any
other family. It is certain also that further species remain to be
discovered, since in addition to the species listed below, young
inflorescences supposed to belong to a _Copernicia_ were collected by
Sintenis (no. 6512) near Utuado, and he also collected two other
_Thrinax_-like palms of doubtful identity, one near Cabo Rojo and one at
Fajardo.


                    Key to the Genera of Sabalaceae

  Leaves depressed in the middle, with a distinct decurved midrib; a
      slender fiber rising from each of the notches which separate the
      leaf segments.

                                                                 INODES.

  Leaves flat, midrib rudimentary; segment without alternating fibers.

        Leaves chartaceous, naked on both sides when mature, the
            veinules unequal; fruits nearly sessile; seeds smooth,
            albumen solid except for a deep basal cavity.

                                                                THRINAX.

        Leaves tough and coriaceous, the lower surface silvery with a
            persistent, closely appressed pubescence; veinules equal;
            fruits distinctly pedicellate; seeds deeply grooved or
            furrowed.

              Trunk tapering upward, tall and slender; pedicels short,
                  bracteate at base; seeds subspherical, ruminate with
                  deep narrow grooves; surface with a dull membranous
                  cuticle.

                                                              THRINCOMA.

              Trunk columnar, of equal diameter or enlarged upward;
                  pedicels long, bracteate above the base; seed naked,
                  smooth and shining, cerebriform, the surface irregular
                  with broad furrows and convolutions.

                                                               THRINGIS.


                           =Inodes= gen. nov.

In this genus, of which the hat palm of Puerto Rico may be considered
the type, it is proposed to accommodate the dendroid palms commonly
referred to _Sabal_, the type of which is _S. Adansonii_ Guersent. The
most conspicuous difference between _Inodes_ and _Sabal_ is, of course,
the fact that the former produces an upright trunk while the latter has
only what might be called an underground rootstock, although such a
distinction is quite artificial, both groups of species beginning life
with a creeping axis which becomes erect in one and remains horizontal
in the other. A much more important difference is to be found in the
leaves which in _Inodes_ have secured strength by the development of a
midrib, a tendency early abandoned by _Sabal_ in which the midrib is
rudimentary and the middle of the leaf is the weakest part. The leaves
of _Sabal_ are adapted for standing erect and avoid resistance to the
wind by being split down the middle. The leaves of _Inodes_ which are
held horizontal from an erect axis have attained the unique adaptation
of a decurved midrib which braces the sloping sides of the leaf and
effectively prevents the breaking above the ligule common in some of the
species of _Thrinax_. It is true that leaves of young specimens of
_Inodes_ stand erect like those of _Sabal_ and do not have the curved
midrib, but even at this stage the midrib is relatively well developed
and the blade opens out to an almost circular form instead of occupying
an arc of 180 degrees or less as in the more strictly flabellate leaves
of _Sabal_.

Further differential characters might be enumerated, such as the short
ligule and the flat petiole of _Sabal_. The inflorescence and seeds also
afford differences, but these points are unnecessary for diagnosis, and
their proper expression will require careful comparative study of the
species of both genera, since _Sabal_ is not monotypic but includes at
least two species from the Southern States and perhaps _S. Mexicana_
Martius. Guersent’s _S. Adansonii_, the first binomial species to which
the name _Sabal_ was applied, is, to judge from the figure, the smaller
of our species, while Jacquin’s _Corypha minor_ may be the larger. Both
species were described from hothouse specimens and the plates give no
details really adequate for identification, but if there are but two
species to be considered there can be little doubt that Jacquin’s
drawing represents the larger of the two forms commonly referred to
_Sabal Adansonii_, since the leaves are nearly four feet long with the
mesial divisions united somewhat less than half way up. The basal
segments are represented, however, as diverging horizontally and not
obliquely as is usual in the living plants in the greenhouses of the
Department of Agriculture.

Guersent maintained that he was dealing with the _Sabal_ which Adanson
had in mind in naming the genus, and made his specific name in
accordance with that fact, treating _Corypha minor_ Jacquin, _Corypha
pumila_ Walter and _Chamaerops acaulis_ Michaux as synonyms. The
relative merits of these names and of _Chamaerops glabra_ Miller, which
Dr. Sargent (Silva, =10=: 38) has resurrected, are not likely to be easy
of determination, but since the last was based on plants grown from
seeds which came from Jamaica, it seems unwise to use it for United
States species to which the description is inapplicable. Miller’s name
may, however, replace _Sabal taurina_ Loddiges which was also founded on
a stemless _Sabal_ supposed to come from Jamaica.

The species of _Inodes_ are in a similar or even worse state of
disorder. There is little use, for example, in transferring to the new
genus the traditional name _umbraculifera_ which was based by Martius on
the _Corypha umbraculifera_ of Jacquin, but not on Linnaeus’ species of
the same name, which is a native of Ceylon. Present taxonomic methods
forbid such generic transfers of misapplied names, so that the name
=Inodes Blackburniana= (_Sabal Blackburniana_ Glazebrook, Gardener’s
Mag. =5=: 52. 1829) should be used instead of the traditional _Sabal
umbraculifera_ of the conservatories, though the identity and origin of
the species still remain in doubt.


                      =Inodes causiarum= sp. nov.

  Trunk 45–75 cm. thick at base, 5–15 m. tall, columnar or slightly
  tapering upward; surface narrowly rimose or nearly smooth, light gray
  or nearly white. Leaf-bases splitting into rather brittle fibers,
  partly remaining compacted into long ribbons 5–8 cm. wide. Leaves
  about 4 m. long, the petiole subequal to the blade, considerably
  exceeded in length by the inflorescence. Petiole 3.8 cm. wide,
  distinctly carinate above near the end; ligule 4.2 cm. in diameter.
  Fruit grayish, 9–10 mm. in diameter; seed chestnut-brown, finely
  rugose or nearly smooth, 7–8 mm. in diameter; embryo oblique, at an
  angle of somewhat less than 45 degrees from the horizontal. Type
  specimen from Joyua (no. 154).

The palm-leaf hats manufactured in large quantities in Puerto Rico are
made from the present species. The center of the hat industry is at
Joyua, a small village on the western coast of the island some miles
southwest of Mayaguez and west of Cabo Rojo. Here many hundreds of the
palms are growing along the shore in a narrow belt of coral sand.

From the two species of _Sabal_ recognized by Grisebach _Inodes
causiarum_ differs from _umbraculifera_ in having the inflorescence much
longer than the leaves, while the trunk and leaves are much shorter and
thicker than in _Sabal mauritiiformis_ a native of Trinidad and
Venezuela which appears from Karsten’s figure, reproduced in the
Natürlichen Pflanzenfamilien, to have neither the leaves nor the habit
of an _Inodes_ though there is no other genus to which it can be
referred with greater propriety. The diameter of the trunk of the
Trinidad palm described as _S. mauritiiformis_ is given as from 12 to 15
inches, while _I. causiarum_ is often two feet or more thick.

From the Florida palmetto, =Inodes Palmetto= (_Corypha Palmetto_ Walter,
Fl. Carol. 119. 1788) the Puerto Rico species differs most conspicuously
in not retaining the old leaf-bases which give the trunk of the Florida
palm so rough an appearance. The cause of this difference is doubtless
to be found in the fact that as with most other palms the trunk of _I.
Palmetto_ grows to full size while the surrounding leaf-bases are still
alive, but in the West Indian species the trunk tapers greatly,
especially in young trees, and the leaf-bases are torn away by its
gradual enlargement to full diameter. The existence in southern Florida
of an _Inodes_ having this last characteristic is a fact of much
interest recently brought to my attention by Mr. E. A. Schwarz, of the
U. S. Department of Agriculture. The specific distinctness of this palm
was impressed upon Mr. Schwarz, not only by its naked trunk, different
habit, and smaller size (5 m., instead of 10 to 20 m.), but also by the
possession of a distinctly tropical insect fauna, quite different from
that of the more northern palmetto with which he had previously been
familiar.[2]

This new Florida species it gives me pleasure to name =Inodes Schwarzii=
in honor of its discoverer, in whose opinion of its distinctness I have
great confidence, although he makes no claims to botanical skill. It is
confined, as far as observed by Mr. Schwarz, to the coral reef formation
of southern Florida, the most accessible station visited being about one
mile south of Cocoanut Grove on the coral reef of the mainland side of
Biscayne Bay. In the vicinity of Snapper Creek, _Inodes Schwarzii_
extends to the Everglades where it is met by _I. Palmetto_. It was also
seen on the Perrine Grant about six miles from Cocoanut Grove; it seemed
not to occur about Miami but reappeared with the appropriate formation
and attendant fauna at New River, though again absent at Lake Worth. A
photograph secured by Mr. H. J. Webber (negative 164) on Taby Island
near Long Key shows an _Inodes_ with a naked trunk and a smaller crown
of straighter leaves than are normal for _I. palmetto_. Messrs. Swingle
and Webber had also remarked the distinctness of the smooth-trunked
palmetto of South Florida.

A third robust species of _Inodes_ is growing in the conservatory of the
Department of Agriculture labeled _Sabal umbraculifera_. It differs
conspicuously from _I. causiarum_ by the very large leaves and by the
great development of fine brown fibers which fill all the interstices
between the leaf-bases, and suggest the name =Inodes vestita=.[3]
Photographs of both the species have been prepared for the illustration
of comparative detailed descriptions.

_Sabal Mexicana_ has been reported from Cuba, and as it is described in
Sargent’s Silva (=10=: 43) as having a trunk “often 2½ feet in
diameter,” a robustness equalled only by the Puerto Rico trees, the
question of its identity was examined. It appears that the original of
_S. Mexicana_ came from southern Mexico and is a trunkless or very
slender, rather than a robust species, being only about 10 cm. in
diameter. The berry and the seed are described as closely similar to
those of _Sabal Adansoni_. Sargent’s _S. Mexicana_ from southern Texas,
in addition to the seven times greater thickness of the trunk, has a
seed nearly 1.25 cm. broad with a strongly prominent micropyle. There
can be little doubt that it is another new species, quite distinct from
that of Puerto Rico, similar only in the unusual diameter of the trunk,
which is furthermore described as bright reddish brown instead of white
or very light grayish as _Inodes causiarum_. In the view of the
apparently localized distribution of the species of this genus the name
=Inodes Texana= would be appropriate for that described and figured by
Sargent as noted above.

In addition to the recently described =Inodes Uresana= (_Sabal Uresana_
Trelease, Rep. Mo. Bot. Gard. =12=: 79), there is another large-seeded
_Inodes_ on the western slope of Mexico, a specimen of which was
collected at Acaponeta, State of Tepic (no. 1528) by Dr. J. N. Rose,[4]
for whom this species may be named =Inodes Rosei=. The seeds are of the
same size and shape as those of _I. Uresana_, but have the surface much
more finely rugose, or nearly smooth, with the embryo directly lateral,
not subdorsal. The branches of the inflorescence are slender and but
little over 1 mm. in diameter instead of fusiform and thickened in the
middle to nearly 3 mm. as shown in Professor Trelease’s photographic
illustration.


        THRINAX Linn. f.; Swartz, Prod. Veg. Ind. Occ. 51. 1788

In the genus _Thrinax_ were formerly placed all the West Indian
fan-palms with smooth stems and no midribs, but the gradual discovery of
numerous and diverse species has resulted in propositions for
subdivision and segregation on the part of several botanists. As usual
these new groups have been characterized very inadequately, and that
mostly from the flowers and seeds, and with no attempt at establishing
correlations of habit or other vegetative features without which the
classification is likely to remain formal and artificial, as well as
useless for popular and field study. Possibly no ecological differences
exist among the _Thrinax_-like palms of other regions, but in Puerto
Rico there are, as shown in the discussion of the following genus, two
well-defined types, one of which varies the ordinary short columnar
habit by the possession of a tall slender and flexible trunk which
doubtless enables it to compete in a measure with the rapid growth of
the surrounding vegetation, and which is also obviously adapted for
withstanding the force of the strong winds encountered in the exposed
places apparently preferred by palms of this species.

The type of the genus _Thrinax_ is the Jamaican _T. parviflora_, a tree
3 to 6 metres high with the trunk swollen at base. The leaves are said
to be 30–60 cm. long with rigid lanceolate divisions; the stipes longer
than the leaves, terete-compressed. The spadix is said to be terminal,
nearly erect and 60–90 cm. long. The tree grows in dry maritime
situations in Jamaica and Santo Domingo. It does not appear that the
original specimens of this species have been examined by Sargent or
other recent writers, but it seems reasonable to use the name for the
group of short species with uniform albumen and a basal cavity instead
of a complete perforation. Swartz’s statement regarding the seed “_intus
albus, medio ruber_,” in connection with its context “_nauco osseo
fragile tectus_” might possibly be rendered “white inside, red between”
and might refer to the red coat of the seed rather than to a red center
as commonly inferred. Of course Swartz might have cut his seed
transversely, but if so he would doubtless have discovered and noted the
perforation had one existed. Patrick Brown’s account of the Jamaica
species, cited by Swartz, evidently refers to a palm with the habits of
_T. Ponceana_. On the other hand the “very slender” palm referred to
under this name in the Jamaica Bulletin (=I=: 196. 1894) shows greater
similarity with _Thrincoma_.


                      =Thrinax praeceps= sp. nov.

  Trunk 8–12 cm. in diameter at base, columnar or slightly enlarged
  upward, seldom attaining over 3 or 4 meters in height. The leaf-bases
  split in the middle of the midrib and long remain adherent to the
  trunk. When they finally fall away on older trees a rather rough
  grayish and longitudinally chinked rimose surface is exposed.

  The stalks of large leaves measure 75–80 cm. in length and 1.2–1.5 cm.
  in width. The middle divisions of the leaf are 55 cm. and under in
  length and attain a width of 4.8 cm., and in the middle of large
  leaves are united for more than half their length. Cross-veinules
  numerous, distinct in both surfaces but especially the upper. The
  white pubescence or tomentum which clothes the young leaves and is
  especially abundant on the ligule soon disappears, leaving the under
  side glaucous or slightly pruinose.

This species is described at some length a little later in a comparison
of generic characters under _Thrincoma alta_. The type specimen (no.
850) was collected on the precipitous mountain-side which overhangs the
road between Utuado and Arecibo, a short distance to the northward from
the station where _Thrincoma alta_ was obtained.

What is believed to be the same species was collected in a similar
situation on the side of a mountain overlooking the town and valley of
Lares.


                  =Thrinax Ponceana= sp. nov. Plate 43

  Trunk 5–8 cm. or more in diameter, columnar, or slightly tapering or
  enlarged upward, 1–4 m. high; surface coarsely and irregularly rimose
  longitudinally. Leaf-bases separating into abundant rather loose light
  grayish or brownish fibers. Leaves numerous, large, drooping or
  pendant; petioles 65 mm. long, 1.5–2 cm. wide; segments attaining 75
  cm. in length and 3.5 cm. in width, united for half their length. Seed
  smooth, mahogany-brown, 5 mm. in diameter. Type specimen no. 1005.

This species apparently exists in much larger quantities than any other
yet known from Puerto Rico, being the predominant plant on several
square miles of territory along the range of dry limestone hills which
skirt the southern coast of the island, to the west of Ponce. Many of
the palms are scattered among the taller shrubs and trees wherever there
is sufficient soil and water to permit these to grow and yet not enough
to give them exclusive possession, but on many of the drier and more
sterile higher slopes the advantage is with the palms.

This abundance of living material deserves more careful study than could
be given during a very brief visit to this almost uninhabited part of
the island, but one note of systematic interest was made. Several
species of _Thrinax_, of which _T. Morrisii_ Wendland may serve as an
example, have been described chiefly with reference to the relative size
of the leaf segments and the extent of their separation. If the palms
under observation near Ponce belonged, as was believed, all to one
species, it is not only true that the individual _Thrinax_ passes all
the stages from the narrow and grass-like, almost completely separated
segments of the very young plant, to the more than half united leaf of
the large tree, but it also appears to be true that under unfavorable
conditions a _Thrinax_ may not be able to attain to full maturity of
size and form but may at the same time produce flowers and seeds. In the
narrow chinks and crevices of the bare rocks were very small, stunted
trees, obviously of great age, while but a few feet distant a deeper
fissure might hold vegetable débris and moisture sufficient to nourish
vigorous specimens several times the size of their less fortunate
companions. The stunted trees retain in proportion to their size, but
apparently with little reference to their age, the small deeply divided
leaves of young plants and have short few-branched inflorescences,
another difference of supposed systematic importance.

In _Thrinax Ponceana_ the leaves of well grown trees have the middle
divisions united to about the middle; the smaller the leaves, the more
deeply they are divided. A further correlation with size is that of the
“fullness” of the leaf. The basal sinus is not closed by the overlapping
of the lateral divisions as in some species, but the area is too great
for a plane circle and there are one or more folds, more numerous and
deeper in large leaves. The lateral divisions do not lie in the plane of
the others but project upward or backward nearly at right angles with
the plane of the middle divisions.

The middle divisions of large leaves may measure 75 cm. in length by 3.5
and sometimes nearly 4 cm. in width, while the narrowly grass-like
lateral segment is only .8 cm. wide and about 30 cm. long. The lowest
segment is not divided at the tip but is produced into a slender
hair-like seta, 6 or 8 cm. long, making it nearly as long or longer than
the next segment above.

The normal segments are split at the apex to the distance of from 2 to 8
cm. and the tips are usually markedly divaricate, owing to the fact that
the young leaves of this species suffer two impressions from the bases
of older leaves, one near the middle, the other near the end. The
pressure causes the curvature of the unopened leaves, which in turn
causes them to split apart when the leaf expands.

Old leaves are smooth and glaucous on the lower side, but in the younger
state more or less remains of the delicate appressed hairiness present
on the lower surfaces of the newly opened leaves. The lower surface is
distinctly grayish and glaucous, but under a lens it can be seen that
this appearance is due to the presence of numerous whitish points
(stomata?) among which are scattering brownish spots of larger size, the
nature of which remains a question.

The free stalks of the largest leaves attain 65 cm. in length and are 2
cm. wide near the base, 1.5 cm. near the apex. The cross section is
lenticular above, but the upper surface becomes flat toward the base.

Young unopened leaves are covered near the base, both above and below,
with a scurfy white tomentum and the margin of the ligule has a long
white fringe.

To avoid possible error it seems best to make separate entry of the
following notes on specimens which might be considered quite distinct
from the larger and normally mature form of _Ponceana_, but which
represent, it is believed, merely a somewhat depauperate condition of
that species, although leaves exactly comparable were not brought home
by our party. The specimens in question were collected by Sintenis (no.
3500) on the south coast of the island near Guanica and distributed from
Berlin as “_Thrinax_ n. sp.”

The leaves are characterized by the narrow straight-sided segments which
retain the same width (15 mm. or less) for about 11 cm.; they are united
in the middle of the leaf for about 8 cm. and the apical tapering part
is about the same length. Other species, so far as known, have the
segments much broader, both absolutely and relatively, and the width is
held for a very much smaller proportion of the length.

In addition the midrib is unusually weak, inconspicuous and only
slightly prominent on the lower side. The small fibro-vascular bundles
which compose it are sometimes spread apart so that there is scarcely an
indication of a rib while in other segments of the same leaf, and
especially at the base, the conditions are more normal. The midrib is
sufficiently distinct above, though very small and fine in comparison
with other species.

Lower surface of leaf glabrous or somewhat glaucous, very slightly
puberulous on the depressed veins near the base. Veinlets inconspicuous,
mostly subequal, though 4 or 5 are sometimes a little larger than the
others. Transverse veinlets indistinct below.

Petiole slender, 4 mm. wide, lenticular in cross section; about 2 mm.
thick. Ligule small and weak, short, with a small apical mucro.

Fruits 5 mm. in diameter, olive brown, irregularly rugose-coriaceous on
the outside as though dried from a pulpy condition; exocarp with a
slightly sweetish taste. Seed bright mahogany-brown, darker below,
depressed-globose, with a sublateral raphe; embryo ascending but more
nearly lateral than vertical; conical basal cavity extending somewhat
above the center, nearly filled with a deep red material.

At the time of our visit in July no ripe fruits of _T. Ponceana_ were
found on the trees, but a few picked up from the ground are apparently
indistinguishable from those of Sintenis’ specimen.


                         =Thrincoma= gen. nov.

  Trunk slender, tapering, flexible; wood firm, covered by a smooth hard
  brittle outer shell or bark.

  Leaf-bases long-sheathing, expanded by the separation of the fibers of
  the side opposite the midrib; petiole strongly flattened above the
  base, prominently angled above and below; ligule large and firm,
  produced laterally to support the outer divisions.

  Leaf-divisions narrow, separated below the middle and below the point
  of greatest width; texture firm and coriaceous; veinules subequal,
  close together, cross-veinules obsolete. Lower surface clothed with
  persistent closely appressed hairs, the upper coated with wax when
  young.

  Seeds with few longitudinal grooves, the surface not polished,
  grayish; embryo subapical.

The generic name alludes to the preference of this palm for the summits
of crags and the brows of perpendicular cliffs which abound in the
limestone region of the north side of Puerto Rico.

The tall, slender trunk and other differences between this genus and
_Thrinax_ are probably to be interpreted as ecological adaptations
necessary to enable the present palm to compete with the vegetation
which often surrounds its base, and to withstand the winds to which it
is commonly exposed. The species of _Thrinax_ and other allied genera,
as far as known, have the trunk rigid and columnar, or even enlarged
from the base upwards. When growing solitary and exposed they seldom, if
ever, attain half the height of _Thrincoma_. Usually, however, they are
protected by other vegetation or by growing gregariously in thickets.

_Thrincoma_ might be described as a _Thrinax_ which has adopted habits
of the arecoid genus _Acria_ which grows in similar situations in a
neighboring part of the island. In addition to the smooth, slender, and
flexible trunk _Thrincoma_ makes further provision against the wind in
having fewer, less ample, tougher and more deeply divided leaves and
like the arecoid palms it also drops the old leaves as soon as their
usefulness is past, instead of retaining, like _Thrinax_, a large
pendant cluster of them. The details of these differences are given
below in a comparative note on fresh material of _Thrincoma alta_ and
_Thrinax praeceps_ collected but a short distance apart in the lower
part of the Arecibo valley along the Utuado-Arecibo road. In this region
of jagged mountains, _Thrinax_ seeks shelter against the walls of
perpendicular precipices, while _Thrincoma_ challenges the wind and the
admiration of the traveller by its evident preference for the crags and
pinnacles.


                       =Thrincoma alta= sp. nov.

With but one species known with certainty to belong to the present genus
the separation of generic and specific characters would have little
purpose. Data for a specific description are, however, contained in the
following notes which are retained in their original comparative form as
better illustrating the generic differentiation of _Thrincoma_ and
_Thrinax_, as represented by _Thrinax praeceps_.

The trunk of _Thrincoma_ differs in three adaptive particulars from that
of _Thrinax praeceps_, _Ponceana_ and similar species which are merely
columnar with very short internodes and an irregularly rimose surface,
not smooth and hardened.

1. There are distinct internodes from 3.5 to 5 cm. in length. These
indicate rapid growth and would increase the chances of survival in the
face of competition of quick-growing tropical vegetation.

2. The trunk tapers gradually from a diameter of 9 cm. near the base to
3.5 at the top, and thus possesses considerable flexibility in view of
its great length, 11 meters, _Thrinax praeceps_ and other related types
not exceeding 4 or 5 meters.

3. In order to support the weight and strain of this greater height, the
texture of the wood is extremely hard and firm, especially near the base
of the trunk. Externally it is covered by a smooth shell or bark of very
hard, brittle, dark colored material. The fibers of the interior which
in _Thrinax_ are merely imbedded in a soft pith like those of a
corn-stalk are here thickened and cemented together, as in tall palms of
other groups, into a dense hard wood. In the specimen cut by us all but
a small area of the middle of the trunk was thus hardened, rendering it
extremely heavy. The wood-fibers of _Thrincoma_ are much coarser than
those of _Thrinax_, and there appear to be none of the obliquely radial
threads which are abundant in the wood of _Thrinax Ponceana_.

With reference to methods of leaf-attachment four differences may be
noted:

1. In _Thrinax praeceps_ the leaf-bases split below in the median line
and remain long attached to the trunk. This adaptation is not confined
to the old leaves but appears while the leaves are still very young, or
as soon as they begin to be expanded by the pressure of those above
them. In the tall species such pressure separates the fibers of the
opposite side of the cylinder. The short species has the outside of the
leaf-bases densely tomentose, and the tomentum is especially abundant
along the edges of the split midrib of the young leaf.

2. The ligule of _Thrincoma_ is notably larger than that of _Thrinax_
and continues to lie in the same plane as the blade, and becomes brown
with maturity. In old leaves of _Thrinax_ the ligule stands nearly at a
right angle to the blade and remains green.

3. For leaves of the same size the petioles, not including the sheathing
base, are longer (75–80 cm.) in the short than in the tall species
(60–65 cm.).

The petiole of the short species is of nearly the same width (1.2–1.5
cm.) throughout, while in the other it is distinctly broader at both
ends than in the middle. The enlargement at the ligule is abrupt. The
base widens gradually to about 2 cm. but is much thinner than in the
short species. In the upper part of the petiole the reverse is true, the
cross section of the leaf-stalk of the _Thrincoma_ being almost
diamond-shape, while that of _Thrinax_ is merely lenticular.

4. These differences of proportion of ligule and stalk are obviously
correlated with the different habits of the two species. The shorter and
more robust trunk of the one enables it to withstand the strain of the
relatively limited exposure to the wind. There is also a greater
flexibility in the leaf itself, due to its thinner texture and to the
smaller development of the ligule and adjacent thickened area, so that
the leaves are often split to near the center. The narrow petiole of the
tall species affords greater flexibility in the lateral plane while
strength has been secured by the greater thickness. On the other hand
the thinness of the base of the petiole of _Thrincoma_ reduces
resistance by permitting the petiole to be twisted when the leaf is
opposed to the wind or blown laterally, thus avoiding the strain which
would come upon the more rigid base of the petiole in _Thrinax_.

The more salient differences between the leaf-blades of the two species
may be enumerated as follows:

1. Although the length of the middle segments of the leaves of
_Thrincoma_ are longer (62 cm.) than those of the other (55 cm.) the
apparent size of the latter is much greater because they are fully
expanded while those of _Thrincoma_ remain more or less fan-shaped,
generally opening less than a semicircle. This decreases the lateral
expansion, since the shortest divisions are brought to the sides, and
gives no projection below the ligule where in _Thrinax_ more than one
third of the foliar expanse is located.

2. The leaf segments are much narrower (3.6 cm.) in the tall than in the
short species (4.8 cm.).

3. Practically the difference in width is still greater because the
segments of _Thrincoma_ are never fully expanded but remain deeply
channelled, thus decreasing the area of exposure to the wind and
increasing the rigidity of the leaf.

4. Resistance to the wind is also reduced in the tall species by the
separation of all the segments to more than two-thirds their length,
while in _Thrinax praeceps_ the median segments are united more than
half way up. In the latter, as in the other members of the group, the
separation begins at the point of greatest width of the segment, but as
if to show that the deeply divided leaves of _Thrincoma_ are an
adaptation, the greatest width is located near the longitudinal middle
of the segments, 10 cm. or more above the bottom of the cleft.

5. The texture of the leaf of _Thrincoma_ is thicker and firmer so that
the segments generally remain straight to the tips while in _Thrinax_
they often droop after the leaves have become fully expanded.

6. The color of the leaves of the tall palm is a very dark green while
those of _Thrinax praeceps_ are uniformly of a much lighter, fresher
tint.

7. The veinules of the firm leaves of _Thrincoma_ are more numerous and
closer together than those of _Thrinax_.

8. The veinules are also subequal in size, giving an appearance of
uniform pattern, while in _Thrinax praeceps_ from 3 to 5 of the veinules
of each side of the midrib are distinctly larger than the others, the
larger veinlets being separated by from 3 to 10 smaller ones.

9. In _Thrincoma_ the cross-veinules are scarcely visible to the naked
eye; under a lens they are still obscure, never equalling in size the
smaller of the longitudinal veinules, which they seldom appear to cross.
In _Thrinax praeceps_, on the contrary, the cross-veinules are as large
as the finer longitudinal ones; they are obvious without a lens and give
the fabric of the leaf a peculiar marbled effect on account of the fact
that they are generally oblique or wavy and commonly appear to cross
several of the longitudinal veinules.

10. The margins of the segments are thickened in both species, and on
the upper side there is a groove inside the marginal rib. In the short
species the margin is flat below and does not become decurved in drying.
In the other the thin edge is closely folded under, and on drying the
sides of the segments uniformly roll under, giving the dried leaves of
the two species an appearance even more dissimilar than in the fresh
state.

11. The lower surface of the leaf of _Thrincoma_ has a silvery white
layer of fine closely appressed hairs, all lying parallel to the veins
and forming a continuous covering. The fibers seem not to be attached
merely at one end, but along the side. They are firmly adherent and are
to be removed only by scraping or rubbing; the surface underneath is
deep green like the upper side, but the fibers remain in the grooves
between the veins. In _Thrinax praeceps_ the lower surface of mature
leaves is smooth and glaucous, a comparatively very slight hairy
covering present in young leaves being evanescent, though traces of it
are usually to be found in the deeper basal grooves. The glaucous
appearance is due to the presence of numerous white or hyaline points
arranged in rows (stomata?). The hairiness of one leaf and the glaucous
character of the other are probably to be looked upon as different
adaptations for the same purpose—the reduction of transpiration.

12. The upper surface and the ligule of young leaves of _Thrincoma_ are
covered with a layer of wax in the form of small plates or scales not
present in _Thrinax_.


                          =Thringis= gen. nov.

  Trunk columnar, rimose; wood pithy. Leaves coriaceous with equal
  veinules, silvery below with closely appressed whitish pubescence.
  Fruits distinctly pedicellate, the pedicel with a bract above the
  base. Seed cerebriform, irregular, with wide furrows and convolutions;
  surface smooth and shining. Embryo subapical.

The characters of this genus are imperfectly known, none of the
specimens being complete. Supposing however, that the association is a
natural one, we have a genus with leaves and pedicellate fruits much
more similar to those of _Thrincoma_ than to those of _Thrinax_, and at
the same time a columnar, rimose and pithy trunk like that of _Thrinax_
and _Coccothrinax_. The seeds appear to differ from those of all related
genera in the possession of large irregular convolutions. The coriaceous
leaves, small fruits, subapical embryo, and other differences separate
this genus from _Coccothrinax_.


                        =Thringis laxa= sp. nov.

  The trunk is columnar or somewhat enlarged upward, about 3.6 m. high
  and 12 cm. in diameter. Surrounding its base was a dense turf of fine
  upright rootlets. The bark was rough and rimose.

  The leaves are similar to those of _T. latifrons_, but smaller, the
  segments being about 70 cm. long by 33 mm. wide. The size of leaves is
  thus about the same as those of _Thrincoma alta_, but the texture is
  thin and flexible, the veinules being slender and not prominent on
  either side. The pubescence is much thinner than that of _T. alta_ and
  of a silvery-gray color.

A palm collected in December, 1899, at Vega Baja, but without fruit (no.
1041). The habit and trunk are not those of _Thrincoma_, but the form
and texture of the leaves and ligule associate the species with
_Thrincoma alta_ rather than with the palms here placed in _Thrinax_.

The columnar habit and protected habitat are reflected in the small
ligule, 18 mm. across, and the relatively broad petiole, 13 mm. wide. It
appears from the dried specimens of this species and _T. latifrons_ that
the leaves may have been “full,” or irregularly folded, instead of
strictly and equally expanded as in _Thrincoma alta_, and the greater
width of the segments is a further indication of this possibility. The
rigidity of the leaf of _Thrincoma alta_ can be maintained because the
segments are narrow and do not open widely.

The soft texture of the leaves of this palm is recognized by the natives
who use it for making hats and call it “yaray” the same name which is
applied in this part of the island to _Inodes causiarum_.


                     =Thringis latifrons= sp. nov.

The leaves, inflorescence and young plants of a palm collected by
Sintenis (no. 3278) on Monte Calabaza near Coamo are much larger and
coarser than those of _Thrincoma alta_. The total length of the middle
segments of the leaf would be over a meter, and the width of the larger
divisions is over 5 cm. The thickness of the petiole at the base of the
ligule is over 10 mm. The form of the ligule is much like that of
_Thrincoma alta_, though scarcely as large in proportion to the size of
the leaf.

The lower surface is clothed with a satiny, appressed grayish pubescence
somewhat less pronounced than that of _Thrincoma alta_. As in that
species the veinules are of equal size, but they are more widely
separated, and the wavy and usually somewhat oblique transverse veinules
are easily distinguishable on both sides of the dried leaf. There are
also slight traces of wax on the ligule and in the grooves of the upper
surface. The median divisions are united for distinctly more than
one-third their length.

The spathes and spadix are distinctly larger than those of _Thrincoma
alta_, but the fruits are, unfortunately, quite immature and contain
only shriveled seeds. The pedicels of the fruits are 2–4 mm. long and
bear, usually near the middle, a very slender bract 1–2 mm. long.

This species is apparently distinct from _Thringis laxa_ in the larger
size and firmer texture of the leaves. It differs in the longer pedicels
of the fruits, with their longer and more slender bracts, from a
specimen belonging to the New York Botanical Garden and supposed to have
been collected by Mr. A. A. Heller, though the number (3278) indicates
that it may belong to the Sintenis series.

This consists of a single, short, once-branched inflorescence arising
from two fibrous spathes. The fruits are about 4 mm. in diameter, nearly
spherical, distinctly apiculate, deep reddish brown in color and borne
on pedicels 2–3 mm. long, with a bract 1 mm. long or less at or below
the middle. The seeds are 2–2.5 mm. in diameter; the surface is smooth
and shining and light brown in color; general shape spherical but with
deep folds and convolutions.

No leaves are known in connection with this specimen, and the exact
locality is also in doubt. Mr. Heller believes, however, that the
inflorescence came from a small _Thrinax_-like palm growing in the
limestone hills a few miles to the east of San Juan.


                            Family ARECACEAE

A large family, with abundant genera in the tropics of America and Asia,
but absent from tropical Africa. The Puerto Rico representatives may be
recognized very easily by the fact that the leaf crown is supported upon
a column of the sheathing bases, a character of which the royal palm
furnishes a conspicuous and ever-present example. Of the remaining
genera, one, the betel palm of the East Indies is sparingly introduced
about towns in the western part of the island and may be recognized at a
glance by reason of the extremely dark green of its foliage. The other
two genera are native palms confined to uncultivated areas and thus
seldom seen at close range from traveled roads. The mountain palm,
_Acrista_, covers the summits of many of the mountains of the island,
but _Aeria_ seems to be confined to the range of high limestone crags
which skirt the northern coast of the island between Bayamon and
Arecibo.


                     Key to the Genera of Arecaceae

  Trunk tall and slender, tapering from a swollen base; spathes numerous
      (7); inflorescence appearing in the axis of the rather persistent
      lower leaves, long and slender; staminate flowers arranged in
      rows.

                                                                  AERIA.

  Trunk robust or of uniform diameter; spathes 1 or 2; inflorescence
      short and brush-like, not exposed until the enclosing leaf below
      it falls away; flowers not set in rows.

        Spathe single, the fruits 2.5 cm. long; leaf-divisions upright,
            very dark green.

                                                                  ARECA.

        Spathes 2, fruits less than 1.25 cm. long; leaf-divisions
            horizontal or oblique.

              Trunk robust, thickened near the middle; leaf-divisions
                  inserted by twos and standing at different angles;
                  inflorescence twice or thrice branched, standing close
                  to the leaf-bases.

                                                              ROYSTONEA.

              Trunk slender, of uniform diameter; leaf-divisions at
                  equal distances, horizontal; inflorescence
                  once-branched, at maturity 15 cm. or more below the
                  leaf-bases.

                                                                ACRISTA.


                           =Aeria= gen. nov.

A tall slender palm evidently related to _Gaussia_, but the embryo
lateral instead of basal, and the pinnae without basal cushions.

Among palms in Puerto Rico _Aeria_ resembles only _Acrista_, from which
it is readily distinguishable by the very slender habit, the swollen
base of the trunk, the much-branched slender interfoliar inflorescence,
the shorter sheathing bases of the leaves, and the numerous spathes.

The embryo of _Aeria_ is located near the longitudinal middle of the
seed on the side opposite the rudiment of the style, which is here
located at the base of the fruit instead of on the side as in _Acrista_.
The albumen is also uniform, except for a small central cavity and the
outer covering is fleshy rather than fibrous.

The position of the embryo is, perhaps, the most obvious difference
between this genus and _Gaussia_, but there are several other
significant discrepancies. Thus the flowers are arranged 3 or 4 in a
row, very seldom 5 or 6. Three fruits develop from one flower only
exceptionally. The trunk is of more than medium height, and the
inflorescence is in reality infrafoliar, for although the dead
leaf-bases and midribs of the leaves are persistent and support the long
inflorescence, this condition is not comparable to that of the cocoid
and other really interfoliar inflorescences.


                  =Aeria attenuata= sp. nov. Plate 45.

The tallest of Puerto Rico palms, probably attaining 30 metres and
upward. The trunk is supported on a mass of coarse roots with spine-like
projecting rootlets arranged in whorls. The surface of the trunk is
smooth with very faint annular impressions. Near the ground the diameter
is 12 to 15 cm. and increases upward to about 25 cm. at about 3 m. above
the base. Above this swelling the trunk tapers very gradually and in
tall specimens is less than 7 cm. in diameter at the top.

The sheathing leaf-base is only 20 cm. long. The leaves remain attached
long after the rupture of the open side, but no fibers are formed, the
edges of the split side being fringed only with brown membranous shreds.
The petiole is rather short, round and rigid and the rachis is
prominently angled above.

Segments of a rather firm texture and standing in different planes, but
all more or less upright or oblique to the rachis, segments from middle
of leaf 2.3 cm. wide near the base, 3.8 cm. long. The segments are set
very closely together, especially the proximal, and overlap each other
in a succubous manner. Fresh fruits deep orange in color and of an
unsymmetrical oval in shape, 16 mm. by 12 mm., with a firm, fleshy outer
covering 1.6 mm. thick, adherent to the seed, the three persistent
styles remain of the same size and are located at the base of the fruit.

The seed is flattened oval, 11 mm. by 9 mm., with a prominent basal
tubercle (hilum). The surface is brownish with a few shallow impressed
lines, but the albumen is white and uniform. Flowers and ripe fruit were
obtained at Vega Baja in December, 1899; type specimen no. 1040.

The so-called llume palm is a most striking ornament of the rugged
limestone hills from Vega Baja to Manati and Arecibo. At a sufficient
distance the slender trunk is no longer visible and the crown of leaves
appears as if suspended in mid-air, while at closer range it does not
seem possible that so slender a shaft can maintain itself. This very
slenderness with the attending flexibility is however, an element of
strength since it permits the trees to bend before the wind while the
leaves diminish the resistance by straightening out as in the cocoanut.
The hurricane of August, 1899, seemed to have done little damage to
these tallest of Puerto Rico palms, many of which project for more than
half their height above everything standing about them. As the trees of
the rather sparse forest growth of these hills are commonly from 12 to
18 metres tall, the llume palms must often attain upwards of 30 metres.


                 ARECA CATECHU Linn. Sp. Pl. 1189. 1753

In the western end of the island the betel palm of the Malay region has
been sparingly introduced, though the fact does not seem to have been
reported hitherto. A few were seen in gardens about Mayaguez and others
in and near San Sebastian. So far as we were able to learn, the people
do not know the name or nature of this introduced species which is
apparently planted only as an ornament or a curiosity. The form is not
unpleasing, but the extremely deep, sombre green of the foliage seems
almost unnatural and imparts a suggestion of artificiality.

Only photographs and fruits of _Areca_ were secured at San Sebastian,
but Puerto Rico specimens collected by Sintenis (no 5749) at Aguadilla
have already been distributed from the Berlin Botanical Garden with the
label “Palma Spec. Subtrib. Attaleae.”


              ROYSTONEA Cook, Science, II. =12=: 479. 1900

_Oreodoxa_ Martius and more recent authors, not Willdenow.

The history of the generic name _Oreodoxa_ shows that botanical writers
of the last few decades have been in error in removing the two original
species and applying it to another series of similar but not closely
related forms. To avoid further confusion with reference to a name which
by reason of the conspicuous character of the trees has wide use in
popular literature it seems desirable to add the following notes on the
genus _Oreodoxa_ as originally established by Willdenow in the Memoires
de l’Academie Royale, Berlin, 1804, a publication which seems to have
been consulted very seldom, even by writers on palms.

Spathe universal, univalvate; spadix ramose, perianth monophyllous,
tripartite below, the divisions ovate, acute, concave; petals ovate,
acuminate, concave. Filaments six, of the length of the corolla; anthers
oblong, acute. Style tripartite, shorter than the filaments, stigma
acute. Ovule, drupe, and seed globose; drupe succulent, but slightly
fibrous; seed single, cartilaginous, nearly smooth, marked with a
longitudinal sulcus. In the discussion subsequent to the statement of
the above characters, _Oreodoxa_ is said to be distinct from _Bactris_
in the tripartite style and in the absence of the “ordinary three
impressions”; it is distinguished from _Areca_, then supposed to include
_Euterpe_ and species now generally placed in _Oreodoxa_, in the single
spathe, the triple style and the hermaphrodite flowers.

The first species is _Oreodoxa acuminata_, referred by recent authors to
_Euterpe_ but probably constituting a distinct genus. The trunk is
erect, cylindrical, very smooth, and attains a height of from 15 to 18
metres; the “root” throws out suckers at the base of the trunk. The
fronds are pinnate, with opposite or alternate, very long, ensiform,
acuminate pinnae, replicate at base. The strongly convolute young leaves
form a green apex for the trunk, five feet high. Spathes cinereous,
folded in at the base of the leaf-sheaths at the top of the trunk,
univalvate, deciduous; spadix erect, much branched, having the
appearance of a broom.

The heart of the bundle of leaf-bases, about two feet long and three
inches thick is eaten as a salad, with oil and vinegar. It is also
stated that the deciduous boat-shaped spathes serve as reservoirs of
rain-water which is long retained in the cool shade cast by the trees.
Birds and beasts, and human natives as well, are said to be dependent at
times upon the liquid thus stored, since in the regions where the palm
grows there are at times no other means of procuring water. The forests
of the high mountain chain of Buena Vista in the province of Caracas are
the native home of the species. It thus appears that in addition to the
structural differences _Oreodoxa acuminata_ occupies quite a different
place in nature from that of the more thoroughly tropical species
commonly referred to that genus, and the stoloniferous habit also
indicates a different ecology.

The second of the original species of _Oreodoxa_ is now referred to the
genus _Catoblastus_. It is a somewhat smaller tree from 12 to 15 metres
high, with a generally similar habit, and is also stoloniferous, but the
pinnae are broad, cuneiform and praemorse, or irregularly truncate as in
the species generally referred to _Martinezia_. The drupaceous fruit is
grayish and the pulp is only slightly succulent; seed the size of a
pigeon’s egg, its exterior brown, marbled with numerous veins. In the
characters of the spathe the arrangement of the fruit and the edible
quality of the heart of the leaf-cluster, as well as in the formation of
lateral off-shoot this species is said to be similar to the first.

Botanists are not yet agreed upon the methods of dealing with
complications like the present in regard to the names of plants, but it
appears certain that those who do not recognize _Oreodoxa_ as a genus
distinct from those admitted in the more recent works on palms must
associate it either with _Euterpe_ or _Catoblastus_. The latter name it
would in that case replace, being much older. Moreover, unless we are
prepared to disregard Willdenow’s statements concerning the
stoloniferous trunk, the simple spathe and the hermaphrodite flowers, to
say nothing of many minor points of circumstantial evidence, there is no
scientific warrant for applying the name _Oreodoxa_ to the noble
Antillean species with which it has been universally associated.

The dried specimens which Willdenow studied were supplemented by notes
of field observation by a court gardener, who was evidently also a
botanist of some experience, to whom Willdenow refers as his “friend.”
The living colors are described with considerable detail throughout the
entire paper, which renders noteworthy the fact that the spathes are
stated to be cinereous. This is in agreement with species of _Euterpe_
which have membranous spathes, but indicates a wide difference from the
West Indian trees where the spathes are thick and fleshy and remain
vivid green until they open and fall away.

The name _Roystonea_ has been given to this ornament of the Puerto Rico
landscape as a respectful compliment to General Roy Stone, the American
engineer officer who secured the admiration of the people of Puerto Rico
by his fearlessness and conspicuous energy in the Adjuntas road-building
campaign which flanked the line of Spanish defenses, and whose
subsequent interest in the improvement of the island will undoubtedly
affect its future history.


           =Roystonea Borinquena= sp. nov. Plate _45. f. 2_.

  Trunk normally fusiform, 30–60 cm. thick, 12–18 m. high. Leaf segments
  4–4.4 cm. in width. Inflorescence robust, compact, twice branched, the
  branches numerous and coarse, ferruginous, pubescent. Fruits
  long-oval, yellowish brown at maturity. Seeds 8 mm. by 6.3 mm.,
  flattened about the hilum, rounded below; wall of endocarp smooth,
  adherent over a small area.

The royal palm of Puerto Rico differs from that of Cuba in having the
trunk generally shorter, more robust and more distinctly fusiform. The
inflorescence is twice branched, with the branches more densely
clustered, coarser and darker colored than those of the Cuban royal
palm, _Roystonea regia_. They are also covered with a slightly hispid
brown pubescence while Cuban specimens are much smoother and more
pallid. The difference of habit, to judge from photographs of the Cuban
species, is most apparent when the trees have grown in the open, as when
planted in avenues or along roadsides. In Puerto Rico, trees which are
obliged to compete with other vegetation are often tall, slender and
unsymmetrical. The typical form is shown in our photograph (no. 250)
taken in the plaza of Juana Diaz.

Martius gives the width of the pinnae of the Cuban royal palm as from 8
to 12 lines. Cuban specimens show as much as one inch and a quarter,
while others from Porto Rico are half an inch wider (44 mm.) of somewhat
coarser texture and with more widely separated secondary veins. The
fruits of the Puerto Rico palm are a deep yellowish brown when ripe,
while those of the Cuban are said to become violet or bluish black.
According to Martius, the fruits of the Cuban species are 6 lines by 4,
but dried specimens show no such discrepancy of proportions and measure
only about 8.5 mm. by 7.5 mm.

In Puerto Rico the fresh fruits are also much longer than broad, perhaps
even more slender than the figures given for the Cuban; when dry they
still appear somewhat longer and larger than the latter.

The seeds of _Roystonea Borinquena_ differ in several particulars from
those of the Cuban species. In shape they are longer and less spherical,
measuring 8 by 6.3 by 5.5 mm. instead of 7.8 by 7 by 6 mm.; the side
bearing the hilum is much flattened and even slightly concave; the
fibers radiating from the hilum are longer, and the corner between the
hilum and the micropyle is evenly rounded, not sharply squared and
prominent as in _R. regia_. On the back of the seed the smooth inner
wall of the endocarp is closely adherent over a small area, while in
Cuban seeds this wall remains attached over nearly the whole side and is
furthermore distinctly rugose-coriaceous on the surface, and has a
distinct sulcus in the median line.

The royal palm is not only the more conspicuous and characteristic
natural object in most parts of Puerto Rico, but it probably exceeds the
cocoanut in total economic importance. The most useful part is the
_yagua_ or sheathing base of the leaf, with which a large proportion of
the houses of the poorer classes are thatched or sided, or both.

The royal palm is one of the wild species which has been distinctly
advantaged by human interference in natural conditions. It is a general
fact that outside the climbing species palms are not successful in
competing with tropical forest vegetation. Originally the royal palm and
the corozo were probably confined to the more rugged slopes of the lower
limestone hills where they both still retain a foothold in places where
the natural growth seems never to have been cleared away. But the vast
majority of royal palms now in existence in Puerto Rico stand on land
which has been cultivated at one time or another, and where the palms
were able to secure a foothold before the competition of other plants
became too strong.

The discovery of root tubercles on a young plant of this species has
been noted in the introductory statement. These tubercles though small
in size are very numerous upon the smaller roots. In shape they are
mostly oval and symmetrical. The larger are about 2 mm. in length though
our natural-size photograph shows several fusiform or clavate bodies
from 5 to 10 mm. long and as much as 2 mm. thick. The color of the roots
and tubercles is white.

The royal palm of Florida is commonly referred to _Oreodoxa regia_,
though with very doubtful propriety. Apparently on account of its great
size, Cooper (Smithsonian Report 1860: 440. 1861) was inclined to
identify it with _Oreodoxa oleracea_ which had also been reported from
the Bahamas. The inflorescence and seeds collected by Curtis on the
western borders of the everglades (no. 2676) are, however, obviously not
those of _R. oleracea_ but are much more similar to those of _R. regia_.
The branches of the inflorescence are much longer and more lax than
those of the species of Cuba and Puerto Rico, from which they also
differ in the frequent development of tertiary branches, in this respect
resembling _Roystonea oleracea_. The fruits do not resemble those of _R.
oleracea_ but are closely similar to those of the other species though
somewhat smaller and more nearly spherical. Several reliable witnesses
are on record to the effect that the trees are from 28 to 35 metres high
and as much as 45 metres has been claimed, while among the royal palms
of Cuba and Puerto Rico 18 metres is the commonly recognized limit of
size. Mr. C. T. Simpson, of the U. S. National Museum, states that the
palms of southwestern Florida lack the conspicuous bulge so
characteristic in the trunks of the Puerto Ricon trees, and that they
grow almost in reach of tide-water, while the natural habitat of the
Puerto Rico species is evidently the limestone hills. In view of these
differences it seems preferable to treat the Florida royal palm as a
distinct species, for which the name =Roystonea Floridana= is proposed.

Mr. Simpson also informs me that the royal palms seen on the islands off
the coast of Honduras had the size and habit of those of Florida and not
the relatively stunted appearance of those seen by him in Hayti and
Jamaica. This fact is suggestive in connection with the popular idea
that the palms of Florida are to be looked upon as recent arrivals from
Cuba. Instead it seems more reasonable to believe that the royal palm of
Puerto Rico, like the species of _Thrinax_ of that island, is a remnant
of the flora of the time when the limestone hills were keys and hammocks
like those of southern Florida, and relatively poor in vegetation able
to crowd out the palms.


                          =Acrista= gen. nov.

  Trunk slender, of uniform diameter. Pinnae horizontal, appendiculate.
  Inflorescences distinctly infrafoliar; spathes two, the outer short,
  the inner long and slender. Spadix once-branched, the branches coarse,
  tapering. Fruits with stigma lateral, seed deeply ruminate, embryo
  basal.

Related to _Roystonea_, but differing in the more slender habit, the
once-branched inflorescence, the basal embryo, and in having the
leaflets in one plane. The color of the foliage is also considerably
lighter than that of the royal palm so that from a distance the general
appearance suggests the cocoanut rather than the royal palm.

There is also some resemblance between the foliage of _Acrista_ and
_Cocops_, but the absence of sheathing leaf-bases in the latter genus
will enable even young specimens to be separated. Moreover the
leaf-divisions of _Cocops_ are much narrower and those at the end of the
leaf are not so much shortened as in _Acrista_.

Further differences from _Roystonea_ are to be found, such as the much
smaller size and the larger roots, which are tuberculate and inclined to
become superficial like those of the llume palm. The sheathing
leaf-bases are not as long proportionately as in _Roystonea_, and there
is a distinct formation of fibers, although the texture is flimsy. The
outer sheaths do not split off and fall away as promptly as in
_Roystonea_ but several dead ones sometimes hang from about the base of
the crown. Although the sheath is longer than in _Aeria_ the fibers are
much better developed, there being but a few membranous shreds in
_Aeria_, and no distinct fibers at all.

Among the mountains between Cayey and Guayama many summits are covered
with the _palma de sierra_, probably in places which have never been
cleared. A few of the palms follow down the steeper uncultivated
ravines. From a distance the crowns suggest royal palms but a closer
view renders the difference apparent. There is also no suggestion of the
bulging trunk of _Roystonea_. In height the _palma de sierra_ probably
does not exceed the royal palm.

The tips of leaflets of young leaves are connected by two brittle red
strands both of which lie on the mesial face, one along the edge, the
other near the middle. The tips of the leaflets are of the same material
and are sometimes persistent as long corneous appendices like those of
the cultivated _Howea_.

The generic name _Euterpe_ Gaertner, which is commonly applied to a
considerable series of American palms related to the present, was in
reality established for the Malayan genus for which the name
_Calyptrocalyx_ Blume is now in use, _Pinanga silvestris globosa_
Rumphius being cited by both Gaertner and Blume as the original, in the
one case, of _Euterpe globosa_, and in the other of _Calyptrocalyx
spicatus_. The origin and identity of the seed described and figured by
Gaertner have not been established, and seem likely to remain in doubt;
but in describing _Calyptrocalyx_, Blume argued that the generic name
should remain with the seeds studied by Gaertner and declared that these
did not belong to any Malayan species but to some of the arecoid palms
of the Mascarene Islands. This suggestion seems not to have been
disposed of by Martius or others, but the fact that Gaertner’s fruits
showed an apical stigma seems to exclude them from the American group
with which the generic name has been associated.

In making use of the name _Euterpe_ for Brazilian palms Martius cites
Gaertner as author of the genus and states that it is of worldwide
distribution in the tropics. Gaertner’s _E. globosa_ is placed as a
synonym of _E. oleracea_[5] Martius, and Jacquin’s older name _Areca
oleracea_ stands in the same relation to _Euterpe edulis_ Martius, thus
rendering _Euterpe oleracea_ Martius a specific homonym. Subsequently
Martius claims the genus _Euterpe_ for himself and expresses doubt
whether it is the same as that named by Gaertner, while Drude in Engler
and Prantl’s Natürlichen Pflanzenfamilien says “_Euterpe_ Mart. (nicht
Gaertn.).” Martius also admits that the West Indian _Areca oleracea_
Jacquin is distinct from the Brazilian species of _Euterpe_, and
redescribes it under the name _Oreodoxa oleracea_.

A further complication connected with _Acrista_ was brought to light by
finding that specimens collected by Sintenis (no. 1525) in the Luquillo
Mountains in northeastern Puerto Rico and distributed from the Berlin
Botanical Garden as _Oreodoxa oleracea_ belong to the present genus,
together with others collected in Martinique by Hahn (no. 805) and
identified at Paris. With the last, the local name _choux palmiste_ is
given, the same which Jacquin noted in the original description of his
_Areca oleracea_ (Stirp. Am. 278. 1763). Moreover, it can scarcely be
determined from Jacquin’s description whether he was dealing with a
_Roystonea_ or an _Acrista_ or with both, though his claim that his was
the tallest palm of the Antilles might hold the name for the
_Roystonea_.

It might then be argued by some that Miller’s species, _Palma altissima_
constituted a segregate from Jacquin’s _oleracea_ and that the latter
name is available for the _Acrista_ of Martinique, whether identical or
not with that of Puerto Rico. But with a possible doubt between the
_Acrista_ and the _Roystonea_ there can scarcely be a justification for
the use of the same name for a third South American species or a fourth
West Indian.

As a means of decreasing the confusion it may be suggested that as
neither the generic nor the specific name of the Brazilian palm which
Martius called _Euterpe oleracea_ (Hist. Nat. Palm. 2: 29) is available,
the name =Catis Martiana= may be proposed, the generic designation
having reference to the drooping pinnae characteristic of the present
species and several of its South American relatives.


                 =Acrista monticola= sp. nov. Plate 44.

  Trunk smooth, 10 to 15 m. high, perhaps taller, from 12 to 15 cm. in
  diameter, with distinct ring-like leaf-scars and internodes, light
  brownish or appearing grayish with bark lichens.

  Leaves about 2 m. long, the pinnae lanceolate, equally spaced and
  lying nearly horizontal, 55 cm. long and 4 cm. broad; the surface
  light green on both sides, with very close parallel longitudinal
  veinlets, but no visible cross veins. The sheathing bases are
  considerably shorter and generally appear somewhat more robust than in
  _Roystonea_. In protected situations the leaf-bases persist and the
  margins shrivel up and expose a flimsy network of fibers.
  Inflorescences appearing several close together; by the falling of the
  leaves above them they are left several inches below the leaf-bases
  before maturity is attained. Spathes fusiform, long, more slender and
  pointed than in _Roystonea_. Spadix once-branched, 1 m. long, 6 cm. in
  diameter at base, tapering gradually to the apex. Branches 23 cm. long
  and less, the proximal branches longest; at first appressed to the
  rachis, the branches are opened out and held stiffly erect by a fleshy
  turgid cushion on the upper (distal) side of the base of each. The
  branches of the rachis may thus be said to be hinged, and with
  maturity the supporting cushion dries away and allows them to resume a
  direction nearly parallel to that of the rachis.

The dried fruits of _Acrista_ are grayish brown in color and nearly
smooth or somewhat coriaceous in external texture; they measure 11 or 12
mm. in length and are nearly as wide, being slightly oboval in shape.
The outer wall is thin and brittle and covers a more or less distinct
thin layer of amorphous brownish material probably representing the pulp
of the fresh fruit; in the dry state this may adhere either to the outer
wall or to the fibers next inside. Near the base these fibers are
simple, pointed and vertical; about half way up they divide and
anastomose and are, as it were, felted and cemented together to form an
oval sac open below and closed above. The outer fibers are much coarser
than the inner and there are sometimes suggestions of three layers
separated by a dark-brown friable material. A few of the delicate inner
fibers are adnate to the surface of the seed which is otherwise free
from its fibrous covering.

Seed 8.5 mm. by 8 mm., slightly lighter in color than the outside of the
fruit. Surface slightly uneven with obscure veinlike ridges and
impressions of the fibers of the outer covering. The kernel is white,
hard and bony, and deeply ruminate, though this is not apparent from the
outside. The channels are very narrow and often radial and straight;
they penetrate 3 mm. or less. Embryo directly basal; hilum lateral,
somewhat below the level of the stigma; a short raphe extends about half
way to the embryo.


                            Family COCACEAE

The cocoid palms are a distinctly American group, the African oil-palm,
_Elaeis Guineensis_ and the cocoanut being the only outliers of the
family which have been supposed to be indigenous in the Old World. South
America is the center of distribution and is the home of a large
proportion of the two hundred or more species. Only five genera reach
Puerto Rico, and one of these, _Cocos_, was probably not a native of the
island.


                   Key to the Subfamilies of Cocaceae

  Trunks, stems, and midribs beset with sharp spines; seeds foraminate
  at or above the middle.

                                                  Subfamily BACTRIDINAE.

  Trunks and other parts unarmed; seeds foraminate at base.

                                                      Subfamily COCINAE.


                         Subfamily BACTRIDINAE

Some of the numerous South American representatives of this group are
nearly smooth, but the three genera known from Puerto Rico have the
trunks, leaf-bases, midribs and inflorescences beset with sharp black
spines, and are thus readily recognizable.


                    Key to the Genera of Bactridinae

  Trunk small, cespitose; leaves separated by long internodes; foramina
      of seeds apical.

                                                                BACTRIS.

  Trunk medium or large, solitary; leaves crowded together at the
      summit; foramina peripheral.

        Trunk slender; leaf-divisions broad, praemorse-truncate;
            pistillate and staminate flowers intermixed on the
            inflorescence; exocarp fleshy.

                                                                 CURIMA.

        Trunk robust; leaf-divisions narrow, sharp-pointed; pistillate
            flowers below and separate from the staminate; exocarp
            fibrous.

                                                              ACROCOMIA.


            BACTRIS Jacquin, Stirp. Am. 279. _pl. 271._ 1763

The type of this genus, _Bactris minor_ Jacquin, described from the
vicinity of Carthagena, Colombia, is a small spiny palm with creeping
rootstocks. The upright trunks are about an inch thick and twelve feet
high, with long spiny internodes. The fruits are fleshy, purple, and
about the size of a cherry. Several species of _Bactris_ are known from
the West Indies though the generic name has doubtless been applied
rather loosely to all the small spiny cocoid palms.

The two following species of _Bactris_ from Puerto Rico described by
Martius several decades ago seem not to have been secured by recent
collectors unless it be true, as suggested below, that one of them, the
simple-leaved _B. acanthophylla_ applies to a young _Curima_. Of _B.
Pavoniana_ the narrowly grass-like leaf-divisions would be sufficiently
characteristic to separate it at once from all other palms known from
Puerto Rico.


            BACTRIS ACANTHOPHYLLA Martius, Palm. Orbign. 67

“Trunk low, spiny; frond simple, the petiole spiny; blade lanceolate in
young plants, oblong in the adult, cuneate at the base and bifid at
apex, the margin unequally erose, unarmed; rachis and primary veins
spiny on both sides; spines bristle-like, narrowed at base, those of the
petiole black, those of the blades fuscous.”

“In the western part of the island of Puerto Rico, near the village of
Yrurena, in swampy places on the margins of aboriginal forests at an
altitude of 400 feet; collected by Wylder, 1827.” (Martius Hist. Palm.
=3=: 281.)

A specimen to which the above diagnosis would not be inapplicable was
collected by Sintenis in the mountain forests near Maricao (no. 484). It
was distributed from Berlin as a _Martinezia_, together with two other
very young plants and a seed to which one of these was attached.

The seed evidently did not come from a cocoid palm but together with the
young seedlings may belong to _Acrista_. The large spiny plant is
probably a young specimen of _Curima_, and should these suggestions
prove to be correct the specific name _acanthophylla_ must be
transferred to this genus though whether it will replace _colophylla_ or
not is not to be determined until it can be ascertained that the Maricao
species is the same as that here described from Bayamon.


              BACTRIS PAVONIANA Martius, Palm. Orbign. 70

“Frond pinnate, rachis with rather long spines and black bristles:
linear acuminate, about equally distant, the terminal united,
setose-ciliate, glaucous below and with a sparse whitish down.”

“Puerto Rico; Pavon.” (Martius, Hist. Pal. =3=: 282.)

Grisebach has reported this species from Antigua and has redescribed it
as follows, presumably from the Antigua specimens.

“‘Trunk low’; _leaves pinnatisect: segments numerous, grass-like,
linear-acuminate_ or the uppermost broader by cohesion, glaucous and
minutely puberulous or glabrescent beneath, approximate, subequidistant,
reduplicate at the base: _rachis armed with very long black prickles_
and rare bristles, keeled above.—Flowers unknown; leaf segments (in our
specimens, which are cut off, perhaps about the middle of the rachis)
more than 30–jugal, 3‴–6‴ distant, 12″–8″ long, 4‴–2‴ broad, superior
gradually shorter, the uppermost cohering ones sometimes 6‴–8‴ broad:
prickles scattered or clustered, slender, the greatest 2″ long. Hab.
Antigua: _Wullschl._, Blubber valley; [Portorico].” (Grisebach, Fl.
Brit. W. I., 520. 1864.)


                           =Curima= gen. nov.

  Trunk rather slender, internodes armed with scattered slender spines.
  Leaves and inflorescence also spiny, especially on the proximal parts.
  Pinnae numerous, strap-shaped, praemorse-truncate, imperfectly
  separated near the ends of the leaves. Inflorescence rather slender,
  once-branched; pistillate flowers mostly located near the bases of the
  branches. Fruit drupaceous, exocarp fleshy, not fibrous; foramina
  peripheral.

A palm related to _Acrocomia_ and to the genera commonly grouped under
the name _Martinezia_, to which _Aiphanes_ and _Marara_ are generally
referred as synonyms. Reasons why none of these names appears available
for the Puerto Rico species are given below. The characters of the
fruit, with foramina near the middle, seem to indicate that _Curima_ is
not remotely related to _Acrocomia_, from which it differs superficially
in the more slender habit, the truncate or praemorse leaves and the very
long and lax inflorescence.


                 =Curima colophylla= sp. nov. Plate 46.

The solitary trunk rises from a mass of spiny roots somewhat smaller
than those of the llume palm (_Aeria_). Diameter of trunk from 1–1.5
cm., often slightly thinner near the ground, though showing no such
tendency to bulge as appears in _Roystonea_, _Aeria_ and _Acrocomia_.
The surface of the internodes is rather sparingly provided with
needle-like spines smaller and more slender than those of _Acrocomia_.
On old trunks the spines are often more or less completely absent.

Leaves 2.13–2.5 m. long, with from 30 to 40 pairs of strap-shaped
praemorse-truncate divisions shorter and broader as the end of the leaf
is approached, and with a terminal undivided area several inches wide.
There is no apparent tendency toward the arrangement of the
leaf-divisions in clusters as in _Martinezia caryotaefolia_ and other
allied species.

The base, rachis, midribs and even the surfaces of the pinnae are beset
with coarse black or deep red spines which are closely appressed when
young and become erect as soon as the surfaces are exposed, all the
parts except the spines and the upper surfaces of the leaf-division
being covered at first with a light grayish or brownish scurfy coating
which gradually disappears.

The inner spathe is narrowly fusiform and about 1 m. long. It splits to
the level of the outer spathe revealing the spadix and its extremely
spiny peduncle. The flowers are greenish cream colored in mass, paler
and not so yellow as in _Acrocomia_. The pistillate flowers are
relatively very few and located near the base of the simple branches.

The cherry-like fruits are dull orange or brick red with rather dry
fleshy or oily exocarp having a rather mealy though distinctly acid
flavor, but no really unpleasant taste. This fleshy covering is only
very slightly fibrous, and that near the base; the seeds fall off very
easily sometimes leaving the base of the exocarp attached to the
fruiting branch. The nut is about 12 mm. in greatest or transverse
diameter and about 10 mm. high, while the fresh fruit is 14–16 mm.
through and 12 or 13 mm. thick. The surface is deeply and irregularly
pitted and marked with three radially fibrous striate foveolae.

It is perhaps too soon to assert that there is only one species of the
present genus in Puerto Rico. The trees certainly differ considerably in
size though not more than the cocoanut and others. There is also a
noticeable difference in the abundance of spines. Such apparent
variability may, however, be due to age, the older trees tending to
become less densely beset with the brittle black spines which are often
conspicuous on young specimens.

The specimens (no. 878) and photographs on which this genus and species
were based were secured on the limestone hills near the wagon road
between Bayamon and Toa Baja where the present palm is not uncommon.

_Curima_ appeared to be especially abundant about Bayamon but is
probably rather generally distributed in the limestone hills of the
island, perhaps also on other soils. A few trees were seen along the
road between Utuado and Lares, and numerous others between Isolina and
Manati. Sintenis collected specimens of what is apparently the same
species near Juncos and Hato Grande, and at Maricao young specimens
discussed under _Bactris acanthophylla_.

As far as Puerto Rico is concerned, this palm is very easily recognized
by means of the curiously truncate leaf-divisions, the outer margins of
which appear as though accidentally injured or eaten away by
caterpillars. This feature is, however, shared with numerous other West
Indian and South American palms, though apparently only one, the
so-called _grigri_ palm of Martinique can be referred to the present
genus with confidence. For this the name =Curima corallina= (_Martinezia
corallina_ Martius, Hist. Nat. Palm. 3: 284) appears to be correct,
although Martius places Gaertner’s much older _Bactris minima_ as a
synonym for his species. Gaertner, however, was making a second attempt
at renaming Jacquin’s _Bactris minor_, having previously misplaced that
name in connection with a West Indian _Acrocomia_, probably the same to
which Jacquin had already supplied the name _Cocos aculeatus_. Thus it
is possible to treat _Bactris minima_ Gaertner as a synonym of _Bactris
minor_ Jacquin and the restoration of Gaertner’s inappropriate name for
the _Curima_ is thus avoided.

With this preliminary description we may return to the consideration of
the generic names _Martinezia_, _Aiphanes_ and _Marara_ which other
writers have applied to relatives of the present palm or treated as
synonyms. _Martinezia_ was described by Ruiz and Pavon (Prodr. Flor.
Per. et Chil. 148. 1794) for five Peruvian palms, but it was amended by
Martius (Hist. Nat. Palm. =3=: 283) by the removal of all the original
species and the substitution of a new set. Of the original species
studied by Ruiz and Pavon only two, _M. ciliata_ and _M. abrupta_ were
mentioned in connection with the original description of the genus, and
this because they offered exceptions to the generic characters. If these
were to be excluded for this reason from those among which the type is
to be sought, the name _Martinezia_ must go with the subsequently
published _M. ensiformis_, now referred to _Euterpe_[6] or with _M.
lanceolata_ and _M. linearis_, now placed in _Chamaedorea_. If we hold
to the first species, _M. ciliata_, _Martinezia_ is probably a synonym
of _Bactris_. The second species, _M. abrupta_, has escaped Martius and
the Index Kewensis, in which a sixth name _M. interrupta_ is the only
one by Ruiz and Pavon now credited as being a genuine _Martinezia_. Thus
by the method of elimination _Martinezia_ would according to current
classification replace _Chamaedorea_ while by the method of types it
would stand as a synonym of _Bactris_.

The genus _Aiphanes_ was established by Willdenow on _Aiphanes
aculeata_, a spiny palm from the mountains about Caracas. The trunk is
said to be erect, ten meters high, subcylindrical and very spiny. The
leaves are about 1.6 m. long, with four pairs of remote, broad, cuneate,
praemorse pinnae, strongly whitish pubescent on the under side; the
petiole is also beset with spines. Spathe acuminate at both ends,
aculeate on the outside, smooth within, opening longitudinally; spadix
4.5 dm. long, composed of cylindrical spikes placed opposite. Flowers
hermaphrodite; calyx trifid, the divisions acute; petals acuminate;
filaments 6, subulate, anthers rounded, style as long as the stamens,
stigma trifid; drupe globose, the fleshy farinaceous pulp rather
tasteless, though edible; nut hard, of the size of a musket ball,
unilocular, black, furrowed with a large number of grayish grooves, of
which three are always much larger than the others. The kernel is white,
very sweet, and very good to eat. _Aiphanes_ grows in the ravines and
forests of the high mountains of the district of Caucagua, province of
Caracas, Venezuela and requires a fertile, somewhat moist soil. It
flowers and fruits in July.

From the above it appears that _Aiphanes_ is a genus quite different
from _Curima_, approaching some of the South American species of
_Bactris_ much more closely than it resembles the Puerto Rico tree.

The genus _Marara_ was based by Karsten (Linnaea, =28=: 389) on _M.
bicuspidata_ from Colombia, a cespitose palm having a trunk 7 meters
high and 10 cm. in diameter, clothed with black spines 6 to 8 mm. long.
The leaves are 125 cm. long with from 60 to 80 pairs of cuneate pinnules
which measure 3 dm. in length and 15 cm. in width, and are clustered in
sixes or eights. This appears to be a very extreme development of the
leaf-arrangement seen in the cultivated palm commonly called _Martinezia
caryotaefolia_ where the leaflets are distinctly clustered, but by no
means so crowded as must be the case when on the side of a leaf 125 cm.
long are leaflets with an aggregate width of 10–13 m.

The palm commonly cultivated in conservatories as _Martinezia
caryotaefolia_ is obviously allied to _Curima_, perhaps more closely
than to either _Aiphanes_ or _Marara_, but in addition to the clustered
pinnules it has a more slender habit, especially apparent in the long
internodes and the more lax inflorescence. This difference in habit is
also evidently correlated with the fact that the leaf-bases do not
become deeply gibbous and obliquely inclined from the trunk as in
_Curima_ but remain closely sheathing. Moreover, the upper side of the
leaf-stalk which in the Puerto Rico palm is deeply channeled and has
lateral corners sharp or torn into fibers nearly to the insertion of the
lowest pinnae is in the conservatory species nearly cylindrical for a
long distance below the pinnae, and has long spines on the upper side as
well as on the lower. It is as though the ligule were located in
_Curima_ near the insertion of the lowest pinnae while in the other form
it remains close to the trunk, with a cylindrical section intercalated
to reach to where the pinnae begin. Apparently we are dealing with still
another generic group for which the name =Tilmia= would not be
inappropriate in allusion to the shorn and disheveled appearance which
it shares with _Curima_. The species studied are =Tilmia caryotaefolia=
(_Martinezia caryotaefolia_ H.B.K. Nov. Gen. et Sp. =I=: 305. _pl. 699_)
in the National Botanic Garden and =T. disticha= (_Martinezia disticha_
Linden, Cat. 32. 1875).

The seeds of _Tilma caryotaefolia_ are like those of _Curima_, but
considerably larger, rounder, and much smoother. The foramina are
peripheral, but are much smaller and more shallow, those of _Curima_
being surrounded, as it were, by a prominent rim which adds somewhat to
the apparent width of the seed. In both genera the nuts are
unsymmetrical, the side which has the largest foramen being distinctly
larger than the others and in _Curima_ the irregularly pitted sculpture
is coarser.


              ACROCOMIA Martius, Hist. Nat. Palm. =2=: 66

A genus of palms distributed through tropical America from Mexico to
Cuba and Paraguay. All the species are of stocky, compact growth, with a
dense crown of numerous leaves. The trunk and the leaf-stalks are
usually armed with strong, sharp spines, sometimes several inches long.

Although totally different on close inspection this genus has in Puerto
Rico a superficial resemblance to the royal palm, which often deceives
travelers. The similarity lies mostly in the two facts that both the
royal and corozo palms are more robust and stiffly erect than the
cocoanut, and that the leaf-divisions instead of lying horizontal and in
one plane are tilted at different angles to the midrib, thus giving the
foliage seen in the mass a somewhat unkempt appearance in comparison
with the cocoanut.

In distinguishing the corozo palm from the royal palm when seen at a
distance so great that the spines of the one and the columnar green
leaf-sheaths of the other can not be seen, recourse may be had to the
following facts. The leaf crown of the corozo palm is much rounder,
thicker and more compact than that of the royal palm, since it contains
many more leaves, and these persist much longer. The royal palm can also
be known by the unopened leaves which project straight upward like
flag-poles or lightning-rods, while in _Acrocomia_ the leaves open as
they are pushed out and seldom offer a suggestion of the spire-like
effect.


                       =Acrocomia media= sp. nov.

  Trunk 20–30 cm. in diameter near the base, thickened above to 50 cm.
  or less; height commonly about 6–8 m. rarely exceeding 10 m. Surface
  of trunk with slight annular impressions. Internodes armed with
  slender black spines, the larger 10–15 cm. long, mostly confined to
  the lower half of the internodes. Fruit green, becoming yellowish, the
  husk firmly fibrous, inedible; about 35 mm. in diameter, nearly
  spherical in shape, with a distinct apical papilla. Kernel 25 mm. wide
  by 22 mm. long; width of the cavity 18 mm. The type specimen was
  collected near Ponce (photograph no. 255).

The _Acrocomia_ of Puerto Rico seems to differ from _A. aculeata_
(Jacquin) in its robust habit and somewhat bulging trunk, while it is
less stout and less swollen than _A. fusiformis_ (Swartz). The name
_Acrocomia lasiospatha_, although used by Martius and Grisebach has no
warrant for supplanting _fusiformis_ of Swartz, which must be preferred
for the Jamaica species with the thick, swollen trunk.

In Jamaica there seem to be at least two species of _Acrocomia_, the
larger of which is called the “great macaw” palm, and is described as
having a fusiform trunk as thick as a man’s body. What is presumably the
same species occurs in Cuba as shown by a photograph from the vicinity
of La Gloria on the north coast. The greatest diameter of the trunk is
three or four times the thickness near the base. In Puerto Rico no trees
approximating these proportions were observed, the greatest amount of
swelling probably not reaching twice the diameter below. According to
Maza _Acrocomia lasiospatha_ grows wild in Cuba and is known under the
name “coroja de Jamaica.” Swartz described his _Cocos fusiformis_ on the
supposition that it was distinct from the _Cocos aculeatus_ of Jacquin,
from Martinique, by reason of the fusiform trunk. The species was,
nevertheless, reduced by Martius to his South American _Acrocomia
sclerocarpa_, perhaps because the spathe is said to be spiny, a
character probably subject to great variation.

Jacquin’s name _Acrocomia aculeata_ (1763) must, it seems, be used for
the West Indian palm placed by Martius under his _A. sclerocarpa_, which
is to be maintained, if at all, as a South American species. Jacquin
declares that the habit of his tree is similar to that of _Cocos
nucifera_ and _Cocos amara_ (_Syagrus_), and his figure shows a tall
straight trunk tapering slightly upward, with no tendency to bulge. The
spines of the trunk are few and the midribs are aculeate on both sides.
The drawing of the fruit is 37 mm. long by 41 mm. wide and has a broad
conic papilla at apex. As indicated above, such a tree was not noticed
in Puerto Rico where all the corozo palms are distinctly, though
slightly, thicker some distance above the base, though apparently never
equaling _A. fusiformis_ in this respect.


                           Subfamily COCINAE


                      Key to the Genera of Cocinae

  Trunk distinctly ringed, rising from an inclined swollen base; leaves
      numerous, many of the lower drooping or pendant, the divisions
      many and narrow; fruits very large, borne continuously.

                                                                  COCOS.

  Trunk nearly smooth, straight and columnar; leaves fewer, not becoming
      pendant, divisions less numerous and broader; fruits small, borne
      at one time and ripening together.

                                                                 COCOPS.


                COCOS NUCIFERA Linn. Sp. Pl. 1188. 1753

The cocoa-palm is largely confined to the neighborhood of the coast, but
is occasionally planted in small numbers in the interior districts,
though it generally does not thrive in such situations especially on the
north side of the island. On the drier southern slope of Puerto Rico,
which is avoided by the royal palm, the cocoanut seems to thrive better,
when it has once become established. Cocoanuts are mostly gathered while
still green, for the sake of the milk or, as it is there called, the
water (_coco de agua_) a popular beverage wherever obtainable. Although
the local consumption of nuts for this purpose is considerable it is
largely confined to the towns of the coast region. Thus it may be said
that in Puerto Rico the cocoa-palm affords a luxury rather than a
necessity, and that it is exceeded in economic importance by the royal
palm.


                           =Cocops= gen. nov.

In a valley on the road between Lares and San Sebastian several young
palms were noticed with leaves similar to the cocoanut, but smaller and
finer. Finally one mature specimen was found, with both trunk and leaves
strongly suggesting the cocoanut, but much smaller. The leaves are light
green, the leaflets in one plane, and the fibers separating from the
narrow base of the leaf. The fibers are few and flimsy, but like those
of the cocoanut and other South American species of _Cocos_. The palm
stood within a few feet of a small permanent brook, down which the seeds
had evidently been carried and there were several young palms along the
bank. The native living in an adjacent house could give us no name
except _palmilla_, and seemed to think that none was necessary since the
tree does not yield _yagua_ or anything else of use. Its early
extermination is therefore not unlikely.

In the absence of flowers and fruit[7] the relationships of the present
genus cannot be ascertained nor its validity satisfactorily established.
There seems, however, to be no reason for including the species in any
of the genera known from Puerto Rico or other parts of the West Indies,
and to associate it with Central and South American types would be a
still less warrantable procedure.

It is also believed that under the present circumstances the application
of a name is justified by convenience of reference and that this will
also assist in securing the attention of botanical collectors better
than a mere allusion to “an unknown palm which may be new.”


                       =Cocops rivalis= sp. nov.

In diameter the trunk appeared to be about midway between the palma de
sierra (_Acrista_) and the cocoanut, and had the short internodes of the
latter. The leaves, however, probably remain somewhat smaller than those
of _Acrista_ to which they might also be said to have a general
similarity, except at the base where their cocoid proclivities become
obvious. At a little distance _Cocops_ might be overlooked as _Acrista_,
while at shorter range it might be mistaken for a very depauperate
cocoanut. No species of _Cocos_ is, however, known to be native in the
West Indies except the doubtful _Cocos crispus_ H.B.K., from Cuba.

As a species _Cocops rivalis_ may prove to be similar to _Syagrus amara_
(Jacquin), which is reported as far north as Jamaica, but it seems to
have no true generic affinity with _Syagrus cocoides_ Martius, the South
American palm which is the type of its genus. According to Martius _S.
amara_ is 30 cm. in diameter, as large or larger than _Cocos nucifera_
and attains the height of from 20 to 35 meters; _Syagrus cocoides_, on
the other hand, is a small slender palm with a trunk 2.5–3 m. high and
5–7.5 cm. in diameter, and with foliage and habit resembling the slender
and diffuse South American species referred by Martius to _Cocos_, but
very different from _Cocos nucifera_ or from _Cocops_.

A leaf collected by Sintenis (no. 6061) near Camuy and coming from
Berlin labeled _Oreodoxa_, obviously did not originate with an arecoid
palm, but probably belongs with the present species. The region of Camuy
is but a few miles from Lares, but there is much extremely rough and
unoccupied country between, so that the danger of extermination appears
to be somewhat diminished.


                         Explanation of Plates

  PLATE 43. _Thrincoma alta_, top of type specimen (no. 848).

  PLATE 44.. _Thrincoma alta_, part of leaf and seeds, natural size.

  PLATE 45.. _Thrinax Ponceana_, type (no. 1005).

  PLATE 46.. _Acrista monticola_, type (no. 761) collected near
  Adjuntas.

  PLATE 47.. Fig. 1, _Aeria attenuata_. Fig. 2, _Cocops rivalis_ (left)
  and _Roystonea Borinquena_ (right).

  PLATE 48.. _Curima colophylla_, apex of flower-cluster and terminal
  leaf-division, natural size. From type specimen (no. 878).

[Illustration:

  PL. 43.

  THRINCOMA ALTA

  HELIOTYPE CO., BOSTON.
]

[Illustration:

  PL. 44.

  THRINCOMA ALTA

  HELIOTYPE CO., BOSTON.
]

[Illustration:

  PL. 45.

  THRINAX PONCEANA

  HELIOTYPE CO., BOSTON.
]

[Illustration:

  PL. 46.

  ACRISTA MONTICOLA

  HELIOTYPE CO., BOSTON.
]

[Illustration:

  PL. 47.

  AERIA ATTENUATA COCOPS RIVALIS
]

[Illustration:

  PL. 48.

  CURIMA COLOPHYLLA

  HELIOTYPE CO., BOSTON.
]

-----

Footnote 1:

  This spelling and the adjective use of the name in this form are
  editorial corrections.

Footnote 2:

  Of numerous insects distinctive of the more southern palmetto the most
  conspicuous is a longicorn beetle, _Agallissus chamaeropis_ Horn, the
  larvae of which bore in the leaf-bases. The more common _Inodes_ is
  inhabited by the allied genus _Zagymnus_, though another species of
  _Agallissus_ is reported from Texas, where the native _Inodes_ is of
  the smooth-trunked type.

Footnote 3:

  =Inodes vestita= sp. nov. Trunk about 45 cm. thick at base, columnar
  or tapering upward; surface rimose, the chinks commonly 5 mm. wide and
  20 mm. apart. Leaf-bases torn into very numerous, fine, hair-like,
  light reddish-brown fibers, a few much coarser than the others and
  measuring from .6 to 1 mm. in diameter. The epidermis separates into
  delicate membranous shreds, the surface of which is delicately pitted
  and sparsely beset with brownish hairy-margined peltate scales.
  Petiole 10 cm. or upward in width below near where it begins to split,
  4.5 cm. wide at base of ligule; 3 m. long, concave above; blade 2.13
  m. long, 2.50 m. wide, composed of about 60 segments, the apical
  united more than two-thirds their length, the basal for less than
  one-third; apical segments 4.5 cm. wide, deeply divided above, a long
  fiber terminating both the longer and the shorter ribs.

As shown by the rimose bark this species affords a rather extreme
instance of the gradual enlargement of the trunk at a distance from the
growing point. Numerous leaf-bases remain attached to the trunk in the
greenhouse as they would not do in nature, since they are torn loose
except for a few fibers at the extreme sides.

Footnote 4:

Dr. Rose also kindly permits the use of the following field notes and
measurements showing that _Inodes Rosei_ is also a taller and more
slender tree than _I. Uresana_.

  “Trees 6–12 or sometimes even 18 meters high, the long slender naked
  trunk 15–20 cm. in diameter, crowned with a large cluster of leaves;
  petioles 60 cm. or more long, flat on the face, pubescent, but
  becoming glabrate; blade pale green, 8 cm. or more in width, strongly
  keeled, more or less clothed beneath with brown scales on the large
  veins; segments cleft to below the middle, 25 mm. or less wide;
  inflorescence in large branching panicles 60 cm. or more long; fruit
  spherical, 18 mm. in diameter, blackish or dark blue when mature.”

“A very common tree east of Rosario towards Mazatlan, also extending all
the way from Rosario to Acaponeta; especially common on the low hills,
and east of Rosario toward the mountains. This species is of
considerable economic importance, the trunks being used in building
fences, corrals and huts, while the leaves appear as thatch on a
majority of the houses of this region.”

Footnote 5:

Hist. Nat. Palmarum 2: 29.

Footnote 6:

Roemer and Schultes treated _Martinezia_ as a synonym of _Oreodoxa_.

Footnote 7:

That the fruits are small and are ripened at one season, as stated in
the key, was apparent from the size of the seedlings and from other
circumstances which accorded with the testimony of the man whose house
stood within a few rods of the largest tree.

------------------------------------------------------------------------




                           TRANSCRIBER’S NOTES


 1. Silently corrected obvious typographical errors and variations in
    spelling.
 2. Retained archaic, non-standard, and uncertain spellings as printed.
 3. Re-indexed footnotes using numbers and collected together at the end
    of the last chapter.
 4. Enclosed italics font in _underscores_.

*** END OF THE PROJECT GUTENBERG EBOOK 69783 ***