178 4. PROOF OF THE COMPACTNESS THEOREM
and that¢ (r) is nonnegative for r E [O, 1] when K::::; 0 or when K > 0 and
VK l'YI :S 7f. Assuming VK l'YI < 7f if K > 0, we compute
(IJI' ¢-IJI ¢')' = IJI" ¢-IJI ¢"
= ( ::2 IJI + K l'Yl
2
IJI) ¢ 2: 0
for all r E [O, l]. Integrating this from 0 tor gives us
IJI' (r) ¢ (r) - IJ (r)I ¢' (r) 2: IJI' (0) ¢ (0) + IJ (O)I ¢' (0) = 0,
that is, for r E (0, 1],
(4.15) IJI' (r) 2: IJ (r)I ~ (~;.
Hence
IJI !!:... IJI I 2: IJl^2 (1) ¢' (l)
dr r=l ¢ (1)
provided VK l'YI < 7f when K > 0. This proves
d I CSKl"l2 (1) 2
-g (\ly exp;^1 x, Y) = IJI -d IJI 2: / (l) !YI ·
r r=l snKli'l2
CSKI · 12 (1)
Note that snKli'l2^7 (1) is positive either when K :S 0 or when K > 0 and
VKl'Yol<1r/2. D
Recall that a 02 function¢ is (strictly) convex if its Hessian is positive
definite: \7V¢ > 0.
COROLLARY 4.46 (Local convexity of the distance squared function).
Suppose the sectional curvatures of (Mn,g) are bounded above by K. Then
the function f (y) ~ ~d^2 (x, y) is convex for any y E B ( x, 7f / ( 2VR)) not
in the cut locus of x.
PROOF. This follows directly from (4.14).
We also have
D
COROLLARY 4.47 (Convexity of small enough balls). Suppose the sec-
tional curvatures of (Mn, g) are bounded above by K. Then the ball B ( 0, r)
is convex if r :S min { inj 0, 7f / ( 2VR)}.
PROOF. Suppose x, y E B ( 0, r). Let / ( t) be the constant speed mini-
mal geodesic between x and y. We simply need to show that d (I (t), 0) < r
for every t. Consider the function f (z) = ~d (0, z)^2 • By Corollary 4.46, we
have that
Vt dtf d (1 (t)) = (\7^2 f) (d' dt' d1) dt > 0,