Microscopical Society of Victoria. 
93 
successive drops on it till high enough. Then he strengthens it 
by further deposits in the form of buttresses from the bottom 
upwards, and lengthens it, making it curve over as seen in Fig. 
3, till long enough, and again he strengthens it. One rib or 
self-standing rafter being finished, he builds the others in like 
manner (marked a in the figures). These form the skeleton, and 
took about fifteen minutes each to build. He is so much master 
of his business that he makes his beginnings, additions, endings, 
and measurements, without ever turning round to look at his 
work, doing everything apparently by feel, with his back to it. 
Then follows the plastering or roofing, which is done inside, 
beginning at the top, beads of glass being stuck to the underside 
of the ribs, and other beads to them, edge to edge, till the whole 
top is covered in, leaving open spaces at the bottom. (This 
plastering is marked b in the figures.) The building thus far 
occupied about four hours, and then my builder rested, I 
presumed, from want of material. One could scarcely credit that 
the whole substance of the house was in his own body a few hours 
before, especially when told (stranger still) that the difference in 
his size before and after the operation was hardly noticeable ; and 
yet the house was twice as wide as the owner was long, and as 
high as he could raise his tail to reach. That one built no more, 
I suppose because as the leaf was drying he could get no sap to 
make more of the secretion. 
I was still in the dark as to how the loops (marked c in the 
figures) were made, as many of them were far higher than the 
makers could possibly reach, and I had never found one on the 
outside of his house as if making them. Further investigation 
settled this point, as I found another specimen in the act of 
spinning them. The plan was this : he stayed inside, head to 
centre, protruded his spinneret under the edge at an open place, 
stuck the beginning of a thread to the leaf or a rib, and then 
kept spinning and pushing backwards his thread which thus made 
a loop, the first half of which was very slender, the last much 
stouter. When long enough, the thick end was fastened to the 
house. These loops took from five to fifteen minutes each to 
spin, and now and then fell over, or met with others that altered 
their direction, for though the semi-fluid sets almost immediately 
enough to keep its shape, it is pliable and sticky for some time. 
