104 Oliver .—On the Obliteration of 
very close relation with a narrow groove which ran longi- 
tudinally along the middle of the dorsal surface. At some 
places a split seemed to extend quite down to the medullary 
hyphae. What this arrangement may signify I know not, 
unless it be to enable a certain amount of the mucilage 
formed between the hyphae to be discharged to the exterior. 
No mucilage-ducts were developed on the concave dorsal 
surface, but plentifully enough around the convex and right 
up to the tips of the horns of the crescent. With these two 
exceptions the zone of sieve-tubes was continuous, the sieve- 
tubes themselves being arranged in radial rows, each row in 
old stems being made up of as many as io or 12 tubes. In 
almost every species of Macrocystis the c bundle 5 is enclosed 
in a well-developed collenchymatous sheath, some 8 or 10 
layers broad, passing over externally into the ordinary cortical 
parenchyma. 
In the medulla run frequent trumpet-hyphae, formed on 
the same type as in Nereocystis, but, generally speaking, 
more often branched. In all but the youngest stems these 
trumpet-hyphae are callous — the callus being formed appar- 
ently from an alteration of the wall. Fig. 1 is a simple 
hypha with callous plate from Macrocystis luxurians drawn as 
seen in corallin-soda. In Fig. 2 a large trumpet-hypha from 
Macrocystis pyrifera is shown ; here it can be seen clearly 
enough that it is the wall which gives rise to the callus ; Figs. 
3 and 4 from M. pyrifera show an excessively common mode 
of branching of these hyphae— similar to that met with in 
Nereocystis (Fig. 6). Variations of this type occur; as for 
instance that given in Fig. 11 (from the medulla of M.plani- 
caalis ), where the main trunk of the hypha divides just in 
front of a sieve-plate. The callus-formation is shown exceed- 
ingly well in this diagram. Sometimes a branching of the 
most complicated nature occurs ; such a branched system of 
trumpet-hyphae is indicated in Fig. 13 from a specimen of 
M . pyrifera brought from the Cape of Good Hope by Brand 
in 1790. All the sieve-plates in this are callous; the walls of 
the hyphae connecting them are not completely involved in 
