49. SAXITRAGACT.T, Ü 
5em.broad, subentire remotely obseurely mueronate-erenate,  pale-gereenish 
above, minutely  punetate-lepidote, very much paler and  albo-lepidote 
beneath, primary veins 6-7 on ench side, connivent near the margin, 
very senbrous on both surfaces, euspidately acute or obtuse at the apex, 
or obtuse at the extremity, rounded at the base.  Panicles terminal, 15 em. 
long 6 em. broad, pyramidal, or racemose axillary. — Calyx punetate broadly 
eampanulate, obseurely 5-suleate, 4 mm. long, 7 mm.in diameter, 5-lobate, 
lobes broadly triangular 2 mm.long 3mm broad shortly acute. — Petals 5, 
"alvate, angustate, 12 mm. long, 5 mm. broad, acute at the apex, lamellate 
on the margin. Stumens 10, 2-seriate, outer ones longer, 123 mm. long, filaments 
11 mm. long, narrowed, nearly 1 mm. broad, apex 2-dentate, abruptly con- 
tracted above the teeth.  Anthers orbieular, retused at both ends. Styles 5, 
11 mm. long, eapitellate at the apex. 
Deulzie | erenate. Sigp.. et. Zuco. 0. "laiwanensis Maxrw.. Hydrang,| Asia. 
Orient. p. 25. 
Deutzia. scabra Hayara im MarsuM. et Havara, Enum. Pl. TFormos. 
p. 92, (pro parte). 
Har. Ako: Tanashü; Taito: Bokusekikaku, Suibi, Kachinro, Kotosho; 
Giran : Hachirisha. 
'l'he present plant is very near JJ. scabra. ''HUNE. and. D. pulehra. Vrboar ; 
but differs from the former in having dentate stamens and in the number 
of styles which is usually 5; from the latter, in having quite obtuse leaves. 
In Vrpar's plant, the leaves are. usually more acuminate, and very aeute at 
the apex.  Deutzie erenata Srgp. et Zuco. 0. Taiwanensis Maxrw.. written in 
* Maxrw, Hydrang. Asia. Orient. p. 23" may possibly be identical with this 
plant, though I have found many different points between the deseription of 
Maxrwowrcz and that of the present one; dor the JDewufzim credited in 
Formosa by MaxrwoWwICZ, as is considered from the loeality given by him, 
may not be otherwise than the present one, as this is the only plant to be 
found in the lowland of Formosa. "The plant written in the paper above 
cited is not near J). erenata, Sigp. et Zvoc., nor is it referable to a variety of 
that species; but must be a quite different one, so far as the diagnosis is con- 
cerned. — As far as T ean judge from the description of the variety 0. 'afiwenensis, 
