210 DR. O, STAPF ON THE FLORA OF MOUNT KINABALU. 
denudati, superne parce divaricatim ramosi, distiche foliati, adpresse cinereo-strigil- 
losi. Folia sessilia, semiamplexicaulia, ob basin cauli arcte adpressam breviter 
semivaginantia, spathulato-cuneata, leviter emarginata, 7-10 lin. longa, 2-3 lin. lata, 
in margine superne adpresse, basin versus patule ciliata, supra sub apice adpresse 
strigillosa, subtus, costa exceptá, glabra, uninervia. Flores pedicello 1-2 lin. longo, 
gracili, adpresse strigilloso suffulti. Calycis tubus brevis, demum 1 lin. longus ; 
lobi demum 2 lin. longi, setuloso-ciliati, cirea nuculas maturas plus minusve patentes. 
Corolle tubus $ lin. longus; lobi tubo equilongi, rotundati, basi constricti. 
Anthers 4 lin. longe. Stylus $ lin. longus. Nucule vix 1 lin. longe. 
At 11,500 feet (Haviland, 1059); at 11,000 feet, in a small marsh (Low). 
Havilandia comes nearest to Myosotis, particularly to those extreme forms which were 
described as M. antarctica by Sir Joseph Hooker, and M. spathulata, Forst. M. spathu- 
lata is found in New Zealand, M. antarctica in New Zealand, Campbell Island, and in 
Magellan Strait. In both species the flowers are solitary in the axils of normal leaves, 
but sometimes the upper leaves are more or less reduced, thus forming an approach to 
the leafless inflorescence of most species of Myosotis. In Havilandia the upper leaves do 
not, as a rule, bear flowers; but they do so occasionally, and, so far, there would hardly be 
sufficient reason to separate the Bornean plant generically from Myosotis. The two 
New-Zealand species above mentioned, however, have exactly the same nutlets as the 
typical Myosotis of the Northern Hemisphere, This is not so in Havilandia. Here 
they are triquetrous, with an acutely carinate ventral edge, and very blunt lateral edges. 
This is the main character on which the new genus is founded. The prefloration of the 
corolla of Myosotis is stated to be always contorted. If this were really the case, another 
differential character might be found in the fact that it is distinctly imbricate in Havi- 
landia. But I would not lay too much stress upon that. First, the way in which the lobes 
overlap each other is not constant in Havilandia, and then M. spathulata and M. antarctica 
at least are exceptions from the type of estivation in Myosotis. The usual form in Havi- 
landia seems to be exactly the same which is figured by Eichler in the diagram for 
Echium vulgare (* Blithendiagramme,’ 197). But I find also that sometimes the lobes 
which correspond to the lateral lobes of the lower lip of Hehiwm are overlapped on both 
sides, in consequence of which the lobe between the calyx-segments 2 and 4 is free 
on both sides. The flowers of Echium are slightly zygomorph. A very slight irregularity 
may be found sometimes also in Havilandia, but normally the corolla is regular. In 
M. spathulata I found in the few buds I examined the same estivation as in Echium, 
whilst in M. antarctica the other form, described above for Havilandia, seems to be more 
common. 
The new genus approaches, on the other hand, Trigonotis, which also has imbricate 
prefloration, but very characteristic nutlets of a different shape. 
SCROPHULARINEJE. 
.— EUPHRASIA BORNEENSIS, Stapf, n. sp. (Pl XVI. B. 10-16.) Herba perennis, a basi 
o. interdum lignescente multiramosa, 3-10 poll. alta, glabra, vel superne minutissime 
