Here is a log of updates, starting in January 2016, to the list of corrections I maintain for the third edition of Mathematical Methods in the Physical Sciences by Mary L. Boas.

August 27, 2022 (contributed by John Reid)
Page 69: In problems 17 and 18, an arbitrary real constant can be added to the right-hand side.
Page 169: In (12.37), the equation \(\textrm{V} = \begin{pmatrix} \phantom{-}8 & -6 \\ -6 & \phantom{-}9 \end{pmatrix}\) should have the letter V in upright font, not slanted font.
Page 783: The solution to problem 5.46 should exclude the case that \(x=y=0\).
November 17, 2021 (contributed by Scott Webster)
Page 807: In the solution to Problem 14.4.8, the Laurent series valid when \(|z|<1\) should be \( -5 +\frac{25}{6}z -\frac{185}{36}z^2+\cdots \). The printed solution mistakenly has \(175\) instead of \(185\) in the numerator of the coefficient of \(z^2\).
November 9, 2021 (contributed by David Calvis)
Page 374: In Problem 5, the instructions imply that the period \(2l\) is equal to \(1/60\), so the hint should say that the value of \(l\) is \(1/120\), not \(1/60\).
September 13, 2019 (contributed by Edward Price)
Page 41: In Problem 22, for “\(n\) is an even integer” read “\(s\) is an even integer”.
May 20, 2019 (contributed by Yoshinao Hirota)
Page 264: In Example 3, the density should be stated to be constant. The same comment applies to Example 4 on page 265, to Problem 1 on page 267, to Problem 26 on page 270, to Problem 4 in the Miscellaneous Problems on page 273, and to Problem 17 on page 274.
May 10, 2019 (contributed by Yoshinao Hirota)
Page 525: At line -5, the parenthetical remark should say “from (8.15)” instead of “from (8.11)”.
March 29, 2018 (contributed by David Calvis)
Pages 588–591: The discussion of solutions of Bessel’s equation has an implicit assumption that \(p\ge 0\), so \(-p \le 0\).
March 11, 2018 (contributed by Tim Leishman)
Page 604: In Problem 4, change \( \lim\limits_{x\to 0} J_{p}(x) /N_{p}(x)\) (the limit of the quotient) to \( \lim\limits_{x\to 0} J_{p}(x) N_{p}(x)\) (the limit of the product).
December 6, 2017 (contributed by Adelaide Deley)
Page 124: At line -6, for “point to pint” read “point to point”.
November 30, 2017 (contributed by Alec English)
Page 499: At line 5, for “nd” read “and”.
November 30, 2017 (contributed by Tim Leishman)
Page 457: In addition to the previously known correction to Example 5, the notation can be clarified by putting a subscript \(0\) on the variables in lines 1, 4, and 7: namely, \( (x_0,y_0,z_0) = (-1, \sqrt{3}, -2)\) and \( (r_0,\theta_0,z_0) = (2, 2\pi/3, -2)\) and \( (r_0,\theta_0, \phi_0) = (2\sqrt{2}, 3\pi/4, 2\pi/3)\).
May 4, 2017 (contributed by Brandon Morrison)
Page 671: In addition to the previously known correction to the equation in the last paragraph on the page (the equation should be set equal to zero), the symbol \(d\) in the first denominator should be changed to a partial derivative symbol \(\partial\).
September 4, 2016 (contributed by Steven Blake)
Page 602: In the integral on the left-hand side of the displayed equation at the bottom of the page, the notation is confusing and technically wrong. The notation \(d(r/a)\) indicates that the integration variable is \(r/a\) (equivalently \(x\)), so the limits of integration should be \(0\) and \(1\), not \(0\) and \(a\). In the integral on the right-hand side of the equation, the integration variable is explicitly \(r\), so the limits of integration are correctly \(0\) and \(a\).
February 10, 2016 (contributed by Tim Leishman)
Page 458: Two lines after the third displayed equation on the page, delete the closing parenthesis preceding the period. In other words, replace the expression \(-\mathbf{e}_r/r^2)\) by \(-\mathbf{e}_r/r^2\) without the trailing parenthesis.
January 28, 2016 (contributed by Jean-Philippe Suter)
Page 356: In the first displayed equation, the expression \( 2\ln x \big|_0^1 \) should be \( 2\ln x \big|_0^\pi \) (but the conclusion that the integral diverges to infinity is unchanged).
Page 624: In the second line following equation (2.14), the expression that arises when \(y=30\) is actually \(\tfrac{1}{2}e^0 - \tfrac{1}{2}e^0\), not \(e^0-e^0\), but is equal to \(0\) as claimed.
Page 634: In the line of text following equation (4.4), for “are are” read “are”.