In Domain Driven Design when an entity clones itself who adds it to its container?

Question

For example, we have two domain objects: Cell and Body (as in human cell and body).

The Body class is just a collection of Cells, e.g.

class Body
{
    IList<Cell> cells;
    public void AddCell(Cell c) { ... }
    public void RemoveCell(Cell c) { ... }
}

The Cell has a Split method, which internally creates a clone of itself, e.g.

Class Cell
{
    public Cell Split()
    {
        Cell newCell = new Cell();
        // Copy this cell's properties into the new cell.
        return Cell;
    }
}

Now, in DDD when the cell splits should:

1. The cell add the newly created cell to the Body (which would mean that each Cell object held a reference to its containing body)?
2. Or should the service layer which received the intitial user request call Split, collect the returned Cell and add it to the Body? (feels like a more anemic design using controllers rather than domain objects)
3. Or should the Body contain a SplitCell method?

Thanks in advance.

Answer 1

I'd think that the Body would simply call the splitCell() on the Cell. Then the Body can do what it wants with the new Cell - add to itself, consume it, throw it away, whatever. The Body contains the Cell anyway.

Answer 2

In DDD, it often depends on, well, the domain. Here, the example - and thus the domain - seems a bit weird but I think I would go for a SplitCell method on Body.

Although it's not very clear to me what the cell splitting means, and what should trigger this action, I guess the body is responsible for splitting its cells. I would be more comfortable with a Regenerate method, or something like that, on the Body, that splits internal cells by calling the Split method on each ones.

Ok, this example is definitively bizarre...