Quantization of a scalar field is a standard text book example of Feynman's path integral quantization. As my findings on the Relativity Theory show that gravitation must be a scalar field, not a tensor field, it is natural to try this quantization method on Nordstrom's and Newton's scalar gravity. It turns out that Feynman's method has many serious errors. The reader doubting it may check the first error very easily. A literature result in equation (11) claimsto give a Green function G(x,x') to the Klein-Gordon operator Box+m^2. If so, (Box+m^2)G=delta(x-x') and if (Box+m^2)y(x)=h(x), then y=int dx' h(x)G(x,x')dx.We see that when integrating over x' the delta peak picks up the value of h(x) becausedelta(x-x') id not 0 in a single point x'=0. But in (11) there is a delta peak delta(x^2)where x^2=|t-t'|^2-|x-x'|^2 is not zero in a single point, it is zero ina subspace. Other errors in Feynman's method are equally clear and real errors. As expected, quantizing gravitation by this method in Section 6 of this article produces a result thatdoes not look correct.