Result 9.2. x→0limxex−1=1.
This result is used here as a known building block (a full elementary justification needs either the series definition of ex or tools developed in a later chapter, so, in keeping with the book, it is quoted rather than re-derived from first principles).
Result 9.3. x→0limxax−1=logea,a>0.
Proof. Since ax and logax are inverse functions of each other, ax can always be rewritten in terms of the natural exponential as ax=exlogea. Hence
xax−1=xexlogea−1=logea⋅xlogeaexlogea−1.
Substitute y=xlogea; as x→0, y→0 as well, and the second factor becomes exactly yey−1, whose limit as y→0 is 1 by Result 9.2. So the whole expression tends to logea⋅1=logea. ■
Result 9.4. x→0limxlog(1+x)=1.
Proof. Set y=log(1+x), so y→0 as x→0, and 1+x=ey, i.e. x=ey−1. Substituting,
xlog(1+x)=ey−1y=yey−11,
and as y→0 the denominator tends to 1 by Result 9.2, so the whole fraction tends to 11=1. ■
Results 9.5–9.9 (stated without proof, used freely from here on):
9.5: limx→0xsin−1x=19.6: limx→0xtan−1x=1
9.7: limx→∞(1+x1)x exists, and equals e9.8: limx→0(1+x)1/x=e9.9: limx→∞(1+xk)x=ek
The number e that keeps appearing here is a transcendental number — it never satisfies any polynomial equation with rational coefficients, anxn+an−1xn−1+⋯+a1x+a0=0 — which is part of why it cannot be pinned down by ordinary algebra and is instead defined through a limit such as Result 9.7.
Worked techniques (Examples 9.32–9.36).
- Reciprocal substitution into the e-form. For a limit such as limx→0(1+sinx)2cscx, substitute y=1/sinx (so y→∞ as x→0) to rewrite the whole expression as (1+y1)2y — exactly the Result 9.7/9.9 pattern — which evaluates to e2.
- Splitting a (1+⋅)power limit at infinity. For limx→∞(x−2x+2)x, first rewrite the base as 1+x−24, substitute y=4x−2 (so y→∞ as x→∞, and x=4y+2), and split the resulting exponent into a "4y" part — which reduces to the standard e4 form — and a leftover finite-exponent part that itself tends to 1; the two combine to give e4.
- Reducing a trigonometric-power limit to the Result 9.7 pattern. For limx→π/4(1−sin2x)5/2(cosx−sinx)5, use the identity (cosx−sinx)2=1−sin2x to rewrite both numerator and denominator purely in terms of (1−sin2x), collapsing the whole quotient to a single power of that quantity; substituting y=1+sin2x (so y→2 as x→π/4) then finishes the limit by direct substitution. …