Female flowers 1 - 3 per bract ; ovary 3 - 4 cm. long, 1.5 - 2 cm. wide, dorsiventrally compressed, orange-yellow, glabrous ; compound tepal 3.5 cm. long, orange-yellow with green lobes, the lateral lobes oval-oblong about 4 mm. long with a spinelike dorsal appendage of about 2 mm., centre lobe smaller and about equal in size to the accessory teeth which are joined to it to form one 3-lobed member ; free tepal as long as the compound, dorsally thickened, opaque, and orange in colour, laterally hyaline ; staminodes short (about 1 cm.) ; style as long as the perianth with a bright orange clavate stigma.

Male bud in advanced blooming very "open", the bracts persisting unwithered for many days, so that there are finally about 20 open at once, but only the last one or two still with flowers ; closed portion of the bud ovoid, the bracts imbricate at the tip ; bracts oblong, lower ones about 7 cm. long, 3 cm. wide, very firm in texture, bright scarlet on both surfaces, not shining on the outside, with the extreme tip yellow or green.

Males flowers 2 per bract ; compound tepal 3.5 - 4 cm. long bright orange with green tip and lobes, the lateral lobes 3 mm. long with a spine-like dorsal appendage of about 1 mm., the centre lobe shorter and without appendage, joined to the accessory teeth ; free tepal as long as the compound, narrow oblong, about 7 mm. wide, dorsally thick, opaque, and orange in colour, laterally hyaline, obtuse at apex ; stamens nearly as long as the perianth, not exserted.

Fruit oblong, 4 - 5 cm. long crowned by the persistent withered perianth, 2 - 2.5 cm. wide somewhat laterally compressed dorsiventrally, rounded at the sessile base, narrowed to a truncate apex ; pericarp about 1.5 mm. thick, orange-yellow at full maturity with a waxy bloom ; pulp white.

Seeds almost cylindrical, black about 6 mm. long, a little wider at the top than at the base, with a distinct waist marking the base of the perisperm chamber within, 4 mm. in diameter at the waist, the surface marked with longitudinal warty ridges".

(Cheesman 1950 q).

These papers reach the same conclusion but by quite different arguments both of which are open to challenge.  The situation is not fully resolved but in the light of these publications M. coccinea is the only valid name for the red flowered Callimusa in cultivation.  The comments below need to be re-written in the light of the above but do at least highlight the confusion in the literature.

The literature on Musa coccinea H. C. Andrews is a little confusing in respect of its spelling and its relationship with Musa uranoscopos J. de Loureiro.

First the spelling. 

Andrews' original 1797 description gives the spelling as Musa coccinea and most literature, including GRIN, gives Musa coccinea.  However, the name is spelled Musa coccinia at Mobot Tropicos and also by Makino 1979, two normally trustworthy authorities, and at TFRI.  It may be that there is good reason for the difference in spelling but I do not know it.

Second, the synonymy. 

In the European Garden Flora, Argent 1984 states that Musa coccinea is synonymous with Musa uranoscopus and gives Musa coccinea as the accepted name.  In the New RHS Dictionary of Gardening, Huxley 1992, gives the same synonymy but gives Musa uranoscopus as the accepted name.  The two sources one might reasonably turn to for guidance here are contradictory.  Recently the Flora of China reduced Musa uranoscopos (there is a difference in spelling here too) under Musa coccinea.

As noted by Cheesman 1950, "J. G. Baker (1893) gives as a synonym of Musa coccinea Andr., "M. Uranoscopos Lour. Fl. Cochinch. 645 excl. syn. Rumph." and if he is right Uranoscopos would appear to be the correct epithet for this species, since Loureiro's Flora Cochinchinensis antedates Andrews description by seven years."

Kurz was certain that Andrew's and Loureiro's plants were the same.  In his chronological review of Musa, Kurz introduces the plant by writing that "Andrews [ ] published Musa coccinea, based upon Loureiro's M. uranoscopos".   Later, while commenting on the rules of botanical nomenclature Kurz says that they "force" him "to call with Loureiro the present Musa coccinea M. uranoscopos".

There is no doubt about the chronological priority of Musa uranoscopos but was J. G. Baker right about the identity of the plants?  E. D. Merrill 1935 obviously thought so and in his commentary on Loureiro's Flora Cochinchinensis he wrote that "from the local name cited by Loureiro [Anamese chúoi tàu] it appears that his specimens were from Indo-China.  His description is manifestly based on actual specimens. [ ] Loureiro's description [of Musa uranoscopos] applies unmistakably to the species with red bracts cultivated for ornamental purposes and currently known as Musa coccinea Andr. [ ]  The binomial Musa uranoscopos Lour. is valid under all rules for this particular species (M. coccinea Andr.)".

That seems clear enough but Cheesman 1950q deliberately refrains from reducing Musa coccinea because he considers that the Musa of Indo-China are not sufficiently well studied to identify Loureiro's plant with the same confidence as Musa coccinea.  He goes on to comment that "until the species of that region have been critically re-examined we should refrain from substituting a somewhat doubtful identification for a certain one". 

I don't think anyone really knows if Musa uranoscopos Lour. and Musa coccinea Andr. are the same thing or not.  As far as I know that critical re-examination of the Musa of Indo-China has not yet been done and the Musa of that region are still relatively little known.  In the meantime, in the horticultural trade the two names are used, as it were, almost interchangeably for the same species.

Cheesman notes that this is the earliest described of the species he grouped together in section Callimusa but its superficial resemblance to other members of the section is not strong.  Evidence is emerging that Musa coccinea/Musa uranoscopos does not belong in section Callimusa.

The plant was introduced to the U.K. about 1791 (B. M. 1559; L. B. C. 475.) and was awarded an AGM by the RHS as Musa uranoscopus.

Specimens: Tsang no. 23320 in Herb. Mus. Paris.

Images.

There are 3 images of Musa coccinea/Musa uranoscopos.