a) Consider remainders of these squares divided by 3. In the big number which we have generated each square can be represented as k2⋅10n, and the big number is the sum of such terms. If k2≡l(mod3) then k2⋅10n≡l(mod3). It is well-known that k2≡0(mod3) is equivalent to 3∣k, and k2≡1(mod3) is equivalent to 3∤k. There are 669 numbers from 1 to 2009 that are divisible by 3, and 1340 there are not. As 1340≡2(mod3) the remainder of the big number modulo 3 is 2, which is impossible for a perfect square, which completes the proof.
b) Consider remainders of these numbers divided by 9. In the big number which we have generated each number can be represented as k⋅10n, and the big number is the sum of such terms. k≡l(mod9) if and only if k⋅10n≡l(mod9). Thus we need to find the sum of the remainders which equals to the sum of all digits of the big number. If we break the number into the groups of nine 1, 2, …, 9; 10, 11, …, 18; …, 1999, 2000, …, 2007, then the sum of the numbers in each group is divisible by 9. The sum of the remaining numbers 2008 and 2009 is divisible by 3 but is not divisible by 9, which is impossible for a perfect square.