pl 
24 Haakon Schetelig. [No. 8 
form type, though it is. always rarer than the round one. More 
interesting is the development of the square plate. From the fol- 
lowing four figures!) we see how the plate is gradually made 
larger, probably to give å better room for the ornaments placed 
here. Sometimes the extended plate is plain as before, sometimes it 
has preserved å higher middle part as å rudiment of the original 
connexion between the bow and the knob (cfr. fig. 30). The ex- 
tension of the plate seems to be no practical improvement and, 
as the spring-coil always is very short as before, there even arise 
some difficulties respecting the fixing of the knobs upon the axis. 
In the brooches tigs. 27—29, the knobs are placed at the back of the 
plate and partly covered by it, but this arrangement must have been felt 
Pio. 1280 2- 
unsatisfactory, as we see that there were at the same time made 
several experiments to bring about a closer connexion between the 
plate and the knobs. For this purpose the axis is made so long that 
1) Fig. 26: Gjedsfjorden, Herø pgd. Nordland. LorancGE: N. Olds. i B. M. 
p3r Ryan: fig. 248. Fig. 27: Øland. Stockholm Mus. The cliché for 
this figure is kindly afforded by år. 0. ALMGREN, Stockholm. Fig. 28 and 99: 
Nordre Fevang, Sandeherred pgå. Larvik. OC. 7027 and 7028. Ab. 1874, p. 143, 
pl. V, fig. 24 and 25. — OConcluding from such brooches as figs. 27—29 I can 
not agree with dr. HILDEBRAND's opinion as to the reason regarding the ex- 
tension of the plate. The plate is evidently not extended to suite the fastening 
of the knobs. Compare dr. HILDEBRAND 1. c. p. 202. 
