The "memorable mathematics" part of this book treats many interesting things. One is "a simple approach to E=Mc^2". First we substitute the relativistic notion of mass m=m_0/sqrt(1-v^2/c^2) into F=ma=d/dt(mv) to get the relativistic F=ma, which is F=m_0a/(1-v^2/c^2)^(3/2). The work done by the force moving a particle from 0 to x is energy=integral from 0 to x of relativistic force=(change in mass)c^2.
Another topic is rocket propulsion in outer space. Consider a rocket with no forces acting on it. Then mv is constant since d/dt(mv)=ma=F=0. The rocket moves forward by throwing out parts of its mass in the form of exhaust products with velocity -b relative to the ship. Since mv is constant we have mv at t=mv at t+dt, i.e. mv=(m+dm)(v+dv)+(-dm)(v-b), which reduces to dv=-b(dm/m) which we can integrate to get, e.g. the burnout velocity for given initial conditions b and fuel/m.
But the best topics are two Euler classics. First the summation of the reciprocals of the squares. (sin x)/x has the roots pi, -pi, 2pi, -2pi, ..., which suggests that the "infinite polynomial" (sin x)/x=1-x^2/3!+x^4/5!-x^6/6!+... should factor as (1-x^2/pi^2)(1-x^2/4pi^2)(1-x^2/9pi^2)... Multiplying this out and equating coefficients of x^2 we get 1/pi^2+1/4pi^2+1/9pi^2+...=1/3!, so the sum of the reciprocals of the squares is pi^2/6. Also, as a bonus, if we put x=pi/2 in the infinite product for (sin x)/x we get Wallis's infinite product for pi.
Euler's study of the reciprocals of the squares also led him to the zeta function zeta(s)=1+1/2^s+1/3^s+..., which he saw can also be written as a product: sum over all integers of 1/n^s = product over all primes of 1/(1-1/p^s), as we see by expanding each factor on the right hand side as a geometric series and multiplying out the product, which gives the reciprocal of each possible product of primes, to the power s, exactly once, i.e., by unique prime factorisation, the left hand side. This charming formula immediately pays off by yeilding a new proof of the old theorem that there are infinitely many primes: because zeta(1)=1+1/2+1/3+...=infinity we have also zeta(1) = product over all primes of 1/(1-1/p) = infinity, which is clearly possible only if there are infinitely many primes.