350 THE JOURNAL OF BOTANY 
| os arranged in eighteen cohorts, as follows :— Poly- 
arpicea, Ranales, Rheadales, Piperales, Malvales, Ebenales (which, 
‘ontodiua C BBB are regarded as descendants of the la se 
named), Geraniales, Myrtiflore, Rosales a huge cohort containin 
Gnetacea), Umbelliflore, Aneisfiore, Passiflorales (culminating i 
Campanulacee and Composite), Centrosperma, Capr ifoliales, ans Tubi.. i- 
flore, a pet descendants of Sterculiacee and divided into two stg 
—Con pocynacee, including dsclepiadacee, Loganiacee, an 
Bubiacen) ‘on Tubiflore in the usual sense, excluding C Ee phexlacee. 
While there is much that is suggestive in Dr. Ha ane r's system, 
one cannot but feel that he has not always viewed affinities in their 
due proportions. For in stance, noting the fam pees oer mong 
the twenty-six included in the cohort Passiflorale find in 
a, ristolochiace 
Begoniacea, Cucurbitacea, Campant ilacee, and Composite. e ae 
clusion of Violacee and Balsaminacee (including Eachewini ii 
series which contains also Composite suggests a view of weinities 
hig it i oh tient to follow, ma one wonders how such a series 
gain removal of Convolvulacee from 
the neighbourhood of Polemoniacee, Hydrophyllacee, Solanaceae, and 
Boraginacea, all of which are members of the last cohort, Tubiflore, 
to an early position in the scheme under Kbenales requires, we think, 
more justification than has hitherto been given. Of his ¢ ohort 
Rosales, which contains thirty-five orders, the author cake that 
it will bebe pale be divided into several after a more pers 
exami 
ism ane itself that a ones would have done a 
still hate olnhal thing towards furthering the progress of phylo- 
genetic poteny if he had deferred the ont pow a of his scheme until 
h een able to make a fuller examination of available material, 
and could give a diagnosis of the cohorts aud orders. 
NEW RUBIACEZ FROM BRITISH EAST AFRICA. 
By Spencer te M. Moors, F.L.S. 
(Concluded from p. 251.) 
THE 
the Sdiowisiz seem wo 
oo decumbens aes in ~ Trop. Afr. iii. 54.—Hab. 
Pemba River. No. 364. 
a ‘tani Linn. fil. Suppl. Pl.127.—Hab. Samburu. No. 487. 
O. Wiedemannit K. Schum. in Engl. Bot, — XXvVili. 57 
(Wiedenmannii) (e deseript.).—Hab. Sultan Hamoud. No. 653. 
y with Schumann’s description, ae that the 
leaves are still narrower tong’ 0-15 em. broad), and the corollas and 
ole of Kaessner’s mate i have now been examined; 
worthy of n 
