I expcnsc of manufnoturc and material is much loss, i t wollld seem as though it should be adopted, and attention tnrnecl to the weight, friction, shape of snrface, etc. Complaint is made of short, light vanes, that they often rnalre a complete revolution in high winds. This could be obviatecl hy increasing t,he weight, bnt this would not be as satisfactory as increasing the length. I t is very evident, that the sa~rie vane will not answer for both light and heavy winds. I t wonld sceirr as

more »

... gh a long flat vane would do for the higher winds ; and the lighter wincls njay be (leternlined h ?the motion of smolre or a light banner, always being careful to keep the line of sight at right angles t o t,he wind. This questlion is an eirlinently practical one. Experime~lts are lll~ich nce(Ie(l to determine the most satisfactory size of surface, length and weight of va~lc, for winds of cliiferent velocities, to satisfy the conditions first laid down. Since writing the above, it bas been si~pgcstecl to me that thc donbic vane can I)e 80 rcaclily braced, it can be made out of very light ninterial, and hence rnay be mnah lighter t,han the fiat vane. The fallacy here confiist.~ in the implication that. a single vane needs :my bracing at all. Since thcrc is no strain upon a flat vane, as it always turns immcdiabely into the aircurrent, it need not bc very stiff; but it is far otherwise with the double vanc. Hcro the sprcading of the tails t~t once brings s tendency to collapse, to each tail, which increases with tho wind-velocity, and is never absent, being greatest when the vune is in the uir-cnrrent. Each t,ail, then, 111iifit be far stiffer than the single tail, which has no strain at any time. But this is not d l : the nlateriul iisnd in the bracing will add nnrcl~ to the weight, especit~lly with the greater allglcs of the tails. For exarnple : t~l c e the most sensitive vane, where 2i = !)OO and (5 = 45".