Well, if the original mass is undergoing projectile motion, it is simpler to suppose it splits at the apex of the quadratic curve, such that the mass originally has zero velocity.
In the absence of air resistance and gravity, the split masses would have to move in opposite directions to conserve momentum. Momentum was zero originally since the whole mass was stationary, and it can't change in the absence of a force.
With gravity, the original mass would still be stationary at the apex of the quadratic, but momentum in the downwards direction would not be conserved because of gravity. The momentum in the two other directions would be zero for the centre of mass because that has to be conserved since there is no net force in those directions. While the masses are still moving in opposite directions, their downwards motion is increasing because of gravity. Thus one mass would have to initially move upwards and the other downwards, so that the mass with downwards velocity hits the ground first.
Now, in this scenario, the centre of mass had no horizontal* component of velocity, so its path was completely vertical, at 90º to the ground. The two masses, are however assumed to have a velocity component in horizontal opposite directions due to whatever separated them to begin with, which add up to zero. Upon collision, the centre of mass moves in the same direction as the moving piece of mass, making its velocity a bit more horizontal. This decrease the angle it makes to the horizontal, and thus makes it less steep. The larger the horizontal velocity component, the less steep it will be. If the two masses had no horizontal component, then there is no change in the angle, since it will fall straight down anyway.
If the mass originally had a horizontal velocity as well, the centre of mass would have to have the same velocity by conservation of components of momentum. There are two ways for that to happen: either the pieces move in the same direction, at a lower horizontal velocity than the centre of mass, or they move in opposite directions, with at least one piece having more momentum than the centre of mass, so that the higher momentum is counterbalanced by the momentum of the other piece.
*What I mean by horizontal is any direction that lies in the plane perpendicular to the direction of gravity.
If they both move in the same direction, the centre of mass loses horizontal velocity when one piece strikes the ground, but also loses vertical velocity as well. At this point it becomes impossible to tell how the angle will change qualitatively. The angle of the velocity of the centre of mass after the piece strikes the ground to the horizontal is the same as that of the still-falling piece. The same holds true if they were moving in opposite directions.
The argument still holds for when the mass does not split at the apex, though it is not necessary for each piece of mass to have vertical velocities in opposite directions in that case. Besides, you could always change your frame of reference to an inertial one such that the velocity of the mass when it splits is zero. Also, note that whatever the angle is initially, it will become steeper and steeper since it is accelerating downwards.
