WING-VENATION OF THE PLECTOPTEEA. 147 
hollow is preserved, in the case of areas broad enough, by the development of: 
intermediate branches which take the opposite condition to the two branches 
which they separate. These have been considered as interpolated veins in 
the JVlayflies ; i.e., they have been supposed to develop froni the margin 
of the wings inwards, and not to be true branches o£ the main veins at 
all. A study of the fossil Majflies shows, however, that originally there 
were no interpolated veins in this Order at all, and that the discontinuity 
of these veins has been brought about actually by absorption of the basal 
connections with the veins of which they are true branches. 
The primitive system of branching which is shown almost to perfection in 
tlie Permian Mayflies is the type which we may conveniently term the triad. 
This can be understood from text-fia;. 2. V is a convex vein which divides 
Diagram of the triadio system of branching- of main veins. For explanation, see text. 
at its primary fork vf into two equally convex branches V] and Y^. In the 
diverging angle between these two, a concave vein arises from either 
Vi or Vs, its point of origin being not far distad from vf; this secondary 
fork is termed vf. In the illustration before us, the concave vein is made to 
arise from Vi, so' that the two branches of this vein become V]a,( + ), 
Vib ( — ), and V2 ( + ). If, instead of being a convex vein, V were a concave 
vein, then the two branches Vj and V2 would be concave, and the inter- 
mediate vein Vii, would be convex. The two cases can be distinguished by 
calling the triad developed from an originally convex vein a positive triad, 
that from an originally concave vein a negative triad. 
Further branching may take place from any one of the three veins of a 
triad, and usually consists of the development of further triads. In text-fig. 2, 
a further triad is seen developed from Vg, viz. the positive triad Vaa( + ), 
V2a'(— )jand V2b( + )- If a triad had been developed from Vib, it would 
have been a negative triad, and the notation used would have been lVib( — ), 
2Vib( + )jand 3Vib( — ); this notation was chosen to avoid duplication of 
suffixes. 
We can now see what has happened to the radius in Protereisma. This 
vein is convex, and divides first of all into a positive triad, Rj ( + ), R2+3 (~)) 
and 114^6 ( + ). The negative member of this group, R2+3) next develops a 
