[CarrotAndStick]

I am not too sure of that... I mean, an offspring will always have 23 pairs of chromosomes, each pair consisting of one half of the father's and one half of the mother's, except for the 23rd one, which comprises the sex chromosomes. All remaining chromosomes are always 50% from the father, and 50% from the mother, and this constitution will be reflected in each and every somatic(body) cell of the offspring.

[beavus]

Yes, half the father's, but probably not all the same half. It is even possible that the two siblings (father and baby) will share no DNA.

Here is a simple example showing just 2 chromosomes to show you how this happens. Line=chromosome, Char 1=ancester (e.g. Alan, Betty, Donna), Char 2=chromosome pair #.

   Grandpa's Parents
    A1 A2    B1 B2
    A1 A2    B1 B2
      |        |
      \---+----/
          |
Mother Grandpa  Grandma
D1 D2   A1 B1    C1 C2
D1 D2   A1 B1    C1 C2
  |       |        |
  +-------^--------+
  |                |
Baby             Father
D1 B1            A1 C1
D1 B1            A1 C1

Here you see that baby has NONE of its father's DNA even though half its DNA came from its father's father.

So it is a statistical problem, where only the expected amount of DNA can be predicted. The range is actually 0% to 50%, both of which are unlikely (about 1 in 8.4 million).

When you factor in random mutations and mitochondrial genetic material, the expected value goes down even further. When you factor in genetic material shared within a population pool (mother may have some of the same genes as father), the estimate then goes up slightly. But overall, the expected amount will be a bit under 25%, with a sizeable variance.