A Treatise on the Theory of Screws
Forfatter: Sir Robert Stawell Ball
År: 1900
Forlag: The University Press
Sted: Cambride
Sider: 544
UDK: 531.1
Søgning i bogen
Den bedste måde at søge i bogen er ved at downloade PDF'en og søge i den.
Derved får du fremhævet ordene visuelt direkte på billedet af siden.
Digitaliseret bog
Bogens tekst er maskinlæst, så der kan være en del fejl og mangler.
APPENDIX I.
485
equations it must also satisfy the 6 - n linear equations which imply that it belongs
to the w-system.
In other words the co-ordinates satisfy (n - 2) + (6 — >i), i.e. four linear equations.
But this is equivalent to saying that the screw lies on a cylindroid. We have thus
the following result.
In the case when the equation for \ has two equal roots, there must be n— 2
separate and distinct principal screws of inertia and also a cylindroid of which
every screw is a principal Screw of Inertia.
For every value of n from 1 to 6 it is of course known that the celebrated
harmonic determinantal equation of § 86 has n real roots.
But when the question arises as to the possibility of this equation as applied
to our present problem having repeated roots, the several cases must be dis-
criminated. It is to be understood that the body itself is to be of a general type
without having e.g. two of the principal radii of gyration equal. The investigation
relates to the possibility of a system of constraints which, while the body is still of
the most general type, shall permit indeterminateness in the number of principal
Screws of Inertia.
Of course if n = 1 the equation is linear and has but a single root.
n = 2. The equation under certain conditions may have two equal roots.
n = 3. The equation under certain conditions may have two or three equal
roots.
n = 4. The equation under certain conditions may have two equal roots or
three equal roots or four equal roots or two pairs of equal roots.
n = 5. The equation can never have a i’epeated root.
n = 6. The equation can never have a repeated root.
Here comes in the restriction that the body is of a general type, for of course
the last statement could not be true if two of the radii of gyration are equal or if
one of them was zero.
The curious contrast between the two last cases and those for the smaller
values of n may be thus accounted for. The expression for T will contain
terms and the ratios only being considered T will contain
n(n+l) _ (n + 2)(n-l)
distinct parameters. As a rigid body is specified both as to position and character
by 9 co-ordinates, it follows that the coefficients of T are not unrestricted if
(?i. + l) (h—1) greater t}lan 9. But this quantity is greater than 9 for the cases
2
of n = 5 and n = 6.
We may put the matter in another way which will perhaps make it clearer.
I shall take the two cases of n = 4 and n = 5.