Schweikert (J. Math. Phys. 45 (1966), 312-317) showed that for suffici
ently high tensions an exponential spline would have no more changes i
n sign of its second derivative than there were changes in the sign of
successive second differences of its knot sequence. Spath (Computing
4 (1969), 225-233) proved the analogous result for first derivatives,
assuming uniform tension throughout the spline. Later, Pluess (J. Appr
ox. Theory 17 (1976), 86-96) extended Spath's result to the case where
the inter-knot tensions p(i) may not all be the same but tend to infi
nity at the same asymptotic growth rate, in the sense that p(i) is an
element of Theta(p(i)) for all i. This paper extends Pruess's result b
y showing his hypothesis of uniform boundedness of the tensions to be
unnecessary. A corollary is the Fact that for high enough minimum inte
rknot tension, the exponential spline through monotone knots will be a
l(2) monotone curve. In addition, qualitative bounds on the differenc
e in slopes between the interpolating polygon and the exponential spli
ne are developed, which show that Gibbs-like behaviour of the spline's
derivative cannot occur in the neighbourhood of the knots. (C) 1997 A
cademic Press.