One of the potential dangers in eclectic texts is that the mechanisms used for indicating variants may mislead readers into isolating variants that are actually in close relationship with each other (and you do see this a lot even in scholarly writing). Take Hebrews 1.3 in NA27 as an example.
The omission of AUTOU is indicated separately from the addition of DI AUTOU - and in any case they relate to different editorially indented (participial) clauses, which could lead you to think that these are basically separate variants, the origins of which could be described and discussed separately.
But some attention to the witnesses for these variants, or an examination of some of the manuscripts, will help show that they are related.
Lacking AUTOU: P46 0243 6 1739 1881* pc
Having DI (E)AUTOU: P46 0243 1739 1881 (and others)
Take a look at P46 at this point:
P46 doesn't omit AUTOU and then add DI AUTOU. It reads DI AUTOU instead of AUTOU. Nor in fact, based on P46, do we really know whether DI AUTOU is linked with FERWN (the preceding verb) or POIHSAMENOS (the following verb) - there is no indication either way. Either way it is difficult to construe, and suggests an odd change of subject.
Perhaps one could even argue that this reading is the harder reading, and ought to be presented as the A-text. If this is printed as the main text, then you would only have a single variant marker leading to two major alternative readings: a) the difficulty of the change of subject leads to the simple modification to DI EATOU in one textual tradition; and b) the deletion of the tricky DI in the other.