Things could be different elsewhere. But this is how it appears to work here:
Males do not corral fry, although it may look like that. Early fry stay together and relate to each other as a swarm -that's pretty much what keeps them together. The male centers his guarding activity around the swarm, making looping turns as he defends them and the immediate vicinity. You can see this behavior in my Spawn Behavior documentary.
As fry mature into young fingerlings (and rapidly), they become stronger swimmers and venture further afield. Although these "early fingerlings" relate to the male (will follow his mass here and there, when he happens to be close enough) they relate more to each other and, more and more, to their environment. They have developed the ability to react with alarm to larger, esp rapidly moving, objects and they commonly spook en masse (flash) if the male appears quickly, or turns quickly. Their reactions to the male's presence can be polar -they spooking at a quick turn of his tail, and then a moment later follow him a short ways. He tends to leave them in the dust though as he's so much faster and has a wider scope, being involved in keeping the vicinity clear of threats.
What appears to happen is that the young bass (at the "early fingerling" stage) leave the male, not vice-versa. They split up into smaller shoals, and many will aggregate at key locations. Cannibalism is frequent so size grouping begins to appear.
My next documentary -in process- is on development from swim-up fry to adulthood. Speaking of that... I've got to get to work. Got ~30 hours of video (2 days, 4 cameras worth) to make into something meaningful and, hopefully, entertaining.