The following figure illustrates the relationship between span (or spread) and the resulting catch and release oarshaft bow angles when rower size (reach) and handle length remain constant:
There seems to be much importance attached to spread changes in the range of one to two centimeters. The Figure shows that changes of this size have virtually no meaningful effect on the oarshaft sweep angles. A one centimeter spread (2cm span) change produces a change--in an overall sweep of ninety degrees--on the order of 0.6 percent (half a degree).
The effect as modeled here for sculls (span) is fairly rigorous because the two-arm reach outward for the catch is shell-axis symmetrical, i.e., the rower does not lean to or favor one side over the other.
In sweeps (spread), however, the result of a spread change would be virtually unnoticed because there is no lateral restraint on the rower's upper body-- lateral motions of which unpredictably and freely change the effective spread. In fact, a look at any photograph of all hands of a four or an eight shows that, while the outboard hand usually always grips the end of the handle, the inboard hand is all over the place. Every centimeter of shift of the inboard hand changes the effective handle length by half a centimeter and the effective spread by the same amount--ergo, there is little point in agonizing over sweep spread. Do whatever seems most natural and comfortable.
Attention should be concentrated on inboard(Lh), oar length(Lo), and lever ratio((Lo-Lh)/Lo), which really do matter--leaving the choice of span and spread to the rower's individual taste and comfort. See "Rower Strength and Oar Length".
Span and spread have no role to play in the determination of oar lever (or "load") ratio. There may be something to be said for altering spread in conjunction with inboard in order to preserve a handle geometry the rower may have become accustomed to. That is, change inboard and span or spread always by similar amounts and in compensating directions.
There seems to be no need for the individuals of fours or eights to have similar spreads nor does there seem to be any advantage in having equal sweep arcs--each should have his lever ratio adjusted to make the best use of his individual strength. But--do make the catch together.
Download the free EXCEL 4.00 file Rigger Geometry, riggeom.xls. By entering your own measurements it will allow you to see just how the catch geometry is affected.