As the OP correctly points out, the concept of number has been systematically generalized throughout the history of mathematics, successively extending it to the rationals, the irrationals, the imaginary numbers, and infinitesimals. Infinitesimals were uppermost in the mind of Leibniz, Euler, Cauchy, and others until the middle of the 19th century; see for example this recent study.
However, starting about 1870, and through the efforts of the great mathematicians Cantor, Dedekind, Weierstrass, and others, the concept of number was specialized to that of a real number (as well as complex, of course), to the exclusion of infinitesimals. Cantor went as far, in fact, as calling infinitesimals "the cholera bacillus of mathematics", "paper numbers", and even "an abomination". See the seminal study by P. Erhlich.
The trend was eventually reversed in the 1960s, but during the period 1870-1960 the number concept furnishes a good example of a specialisation of the meaning of a concept in the history of mathematics.