Geometric integration by parts
Author(s): Jran Corwin
Source: The Mathematics Teacher , January 2002, Vol. 95, No. 1 (January 2002), p. 56, 58
Published by: National Council of Teachers of Mathematics
Stable URL: [Link]
JSTOR is a not-for-profit service that helps scholars, researchers, and students discover, use, and build upon a wide
range of content in a trusted digital archive. We use information technology and tools to increase productivity and
facilitate new forms of scholarship. For more information about JSTOR, please contact support@[Link].
Your use of the JSTOR archive indicates your acceptance of the Terms & Conditions of Use, available at
[Link]
National Council of Teachers of Mathematics is collaborating with JSTOR to digitize, preserve
and extend access to The Mathematics Teacher
This content downloaded from
[Link] on Thu, 11 Aug 2022 02:28:45 UTC
All use subject to [Link]
is a sequence that can be gener discussed above. Using the partial Integrating both sides and re
ated recursively in at least two fraction decomposition arranging yield
different ways. It can be rewritten
as 1 = 1/4 1/4 j udv = u v - j vdu.
1 + 2 3 + 4 8+9 20+21(*-2)(A+2) *-2 fc+2' This identity can sometimes be
employed to solve for the anti
one can see that the nth partial
P 2 ' 5 ' 12 ' 29 '
sum, with k starting at 3, is
derivative of a function that is
More on triples 49 + 50 119+120 288 + 289
difficult to determine by more
conventional methods. Tb solve for
Angelo DiDomenico and Randy
70 9 169 ' 408 9
Tanner have put together a ter
rific lesson on an alternative way 696 + 697 =1HHHMB) J" tan"1 (x) dx,
(HHH)
we set u = tan'Kx) and set dv =
of generating Pythagorean triples
dx. Then
in "Pythagorean Triples from
Harmonic Sequences" on pages
218-22 of the March 2001 issue
The third, fifth, seventh, and
ninth terms of the sequence are
of the Mathematics Teacher. After the Pythagorean triples
P-W-Ml
yi-6/ W + 6/_
du = - -dx
l + xz
I read the article, I too started to
{(3,4,5), (20,21,29), (119,120,169), Notice that the first four positive and v=x. Substituting these val
play a bit by taking the average terms have no corresponding
of any two consecutive terms of a
(696,697,985),...}, ues into our integration-by-parts
negative term with which to formula yields
harmonic sequence. Using the which are the first four of the
cancel. Similarly, four negative
harmonic sequence NIRTs (the nearly isosceles right
triangles) in which the legs differ
terms at the end have no positive jtan_1(x)dx
terms with which to cancel, but
Jiiii...} by 1. The second, fourth, sixth,
eighth, and beyond terms produce
these terms approach 0 as n
approaches infinity, leaving just
-x tan"1 fx)-1 j+ xx2 dx
a certain set of Titterton triples,
I found these averages to be the four terms predicted by the
which, when taken as the elements theorem.
3 5^| 1 of three-dimensional vectors,
possess magnitudes that are also
In general, the partial fraction
decomposition is
= xtan-1(x)-^ln|l + x2 +C
integral: = xtan~1(x)-hn/l + x2+C.
using the harmonic sequence I [1,2,2] | =3 ! 1/(26) 1/(26) That approach is the conven
Jiiii i ...1 | [8,9,12] | =17
| [49,50,70] j =99
(k-b)(k+b)~ k-b ~ k+b 9
tional, algebraic approach to inte
gration by parts. The first step in
UVio'is'ie' J* and after factoring 1/(26) out of the my geometric method was to
| [288,289,408] | = 577 rewrite y - tan"1 (x) in a form that
I found the averages to be nth partial sum, the first 26 posi
John Patrick Titterton I could easily integrate, x = tany.
tive terms (1,1/2,1/3,1/(26))
JET313@[Link]
|5 11 JL7 29 1 Cold Spring Harbor, NY
lack negative terms with which By considering these two equations
and their graphs, the area bounded
[8'56'l40'208' J' 11724 to cancel. As before, 26 negative
between the first curve and the
terms lack positive terms with
The latter sequence does not which to cancel, but these terms x-axis (what our original integral
consist of Pythagorean pairs in will all approach 0 as n approaches solves for) can obviously be found
the numerators and denominators
unless each numerator is multi
Telescoping series infinity, thus proving the result. indirectly by first solving for the
Matthew Donofrio, a student in area bounded by the curve and
Raymond N. Greenwell
plied by 3: my second-semester calculus by the y-axis, as shown in figure
matrng@[Link]
class, came to me one day with
1 (Corwin). Subtracting the area
Hofstra University
Jl5 33 51 87 1 an observation that he had made Hempstead, NY 11549
bounded by they-axis from the
l8'56'l40'208' J on the basis of several examples rectangle shown should result
of telescoping series: in the area represented by the
The corresponding Pythagorean original integral.
triples are (15,8,17), (33,56,65), ~ i _i 261 Geometric integration
(51,140,149), and (87,208, 217).
k=b+l j=l
by parts jtan(y)dy Jtarr1 (x)dx
And thus, a teacher can work
During a recent calculus BC test,
through DiDomenco and Tanner's we were asked to evaluate the
where b is any positive integer.
approach in class and then ask the
For example, when 6 indefinite
= 2, integral . , H-h
students to apply the verification -2 -1.5 -1 ~.5> .5 1 i1.5 2
-.375
and generalization of my obser J tan"1 (x) dx.
vation, or vice versa, as a special
assignment. And perhaps students
2>-4"4Z./
fc=3 M
Because I had not studied, I
+ -75
--1.25
missed the obvious solution that y = tarr1(x)
+ -1.5
can even devise a procedure of -26 used traditional integration by x = tan(y)
their own. 48' parts. Yet, in the time allotted,
While we are on the subject of
I determined the integral by a Fig. 1 (Corwin)
generating Pythagorean pairs Matthew was wondering whether
geometric interpretation of j\an^(x)dx = xy - \\an{y)dy.
from odd places, the sequence of this result was always true, and
integration by parts.
partial sums associated with the why. After studying the result, I
determined that it was indeed a As a review, integration by
continued fraction expansion of Solving for Jtan (y) dy, we find
theorem that I had never seen, so parts comes from the multiplica that
2, that is, tion rule for differentiation. If we
I label it Donofrio's theorem.
consider the functions u and v,
1 3 7 17 41 99 239 577 1393 1 To verify Donofrio's theorem, /tan"1 (x)dx = xy - In | sec(jy) | + C.
\ Y5'i2'29'7o'i69'408'985' \ first consider the case of b - 2, d(u v) = udv + vdu. (Continued on page 58)
56 MATHEMATICS TEACHER
This content downloaded from
[Link] on Thu, 11 Aug 2022 02:28:45 UTC
All use subject to [Link]
(Continued from page 5$) I
journals
Readers of the Mathematics
Then substituting y = tan 1 (x)
results in Teacher might enjoy the fol
lowing articles in the January
2002 issue of Mathematics
/tan"1 (x)dx = x tan"1 (x)
Teaching in the Middle
-lnlsecttan^Oc))! + C. School:
Using right-triangle trigonometry, "Probability on a Budget,"
we find that sec (tan-1 (x)) = VI + x2.
by William A. Ewbank and
Therefore, we arrive at the John L. Ginther
desired formula,
"Fibonacci: Beautiful
Patterns, Beautiful Mathe
J tan"1 (x)dx matics," by Catherine B.
= xtan-1 (x) - In VTT?+ C. Miller and Tamara B.
Veenstra
This same method should be
applied only when the function to For a complete listing of the
be integrated has an inverse that contents of all the journals,
seethe NCTM Web site at
is more easily integrated. The
real reason for teaching this [Link].
method is to give students a
better geometric understanding
of integration by parts.
Ivan Corwin
Arlington High School
LaGrangeville, NY 12540
DIVERSITY IN MATHEMATICS EDUCATION:
CENTER ON LEARNING AND TEACHING
Doctoral and Postdoctoral Fellowships
The University of Wisconsin?Madison, University of To receive a fellowship from an institution, an applicant
California at Berkeley, and University of California at must be accepted to that institution's program. Fellowship
Los Angeles are pleased to announce the availability of sub offers for fall 2002 will be made beginning February 15, 2002
stantial long-term support for doctoral and post-doctoral and will continue until all fellowships have been awarded and
fellows who will pursue studies and research on the mathe accepted. To insure that an application receives the fullest
matics education of diverse learning populations, along the consideration possible by an individual campus, the com
lines of race/ethnicity, class, gender, and language. Under a plete application should be submitted to that campus by no
grant from the National Science Foundation, we will accept later than January 15, 2002.
two cohorts of fellows to begin their work in the fall of Competition of National Science Foundation
2002 and 2003, respectively. Each campus will appoint five fellowships is open to United States citizens or permanent
fellows per cohort (for a grand total of 30 fellows). residents at the time of application.
For further information, please contact one of the collaborating institutions:
Professor Thomas P. Carpenter, Associate Professor Megan L. Franke, Professor Rogers Hall,
Associate Director, DiME/CLT Associate Director, DiME/CLT Associate Director, DiME/CLT
c/o Chris Kruger c/o Cheryl Batiste c/o CD Area Assistant Kate Capps
Graduate School of Education Graduate School of Education
Department of Curriculum & Instruction
and Information Sciences 4533 Tolman Hall
University of Wisconsin?Madison
225 North Mills Street University of California, Los Angeles University of California, Berkeley
Madison, Wl 53706-1795 Los Angeles, CA 90095-1521 Berkeley, CA 94720-1670
Phone: 608-263-5141 Phone:310-825-4910 phone: 510-642 4207
e-mails: tpcarpen@[Link], e-mail: mfranke@[Link] e-mails: rhall@[Link],
hrugei@[Link] kate@[Link]
DiME/CLT WEB SITE: [Link]/dime
58 MATHEMATICS TEACHER
This content downloaded from
[Link] on Thu, 11 Aug 2022 02:28:45 UTC
All use subject to [Link]