Lanes 3 to 7 shown high amounts of DNA in the bands.
In lanes 4,5,7 the inserts (C) have a size just slightly bigger then the vector part.
The vector parts of 3-7 (A) have the same migration rate as the fragment of pGT4 (B).
Note that the smaller BamHI/EcoRI fragment of pGT4 has run from the gel.
The inserts of recombinant clones 4,3 and 1 (C) must be one of the two in this band (D) in the marker lane 8. This bands contains the 3758 bp and the 3775 bp fragments of Lambda DNA.
The 3530 and 5505 bp cos-end containing Lambda DNA fragments (G) are present in low amounts because some of those fragments have joined to form the cos-site containing 9035 fragment (F).

It's clear, that the inserts of clones 2 and 5 (in lanes 6 and 3, see H) are identical to the 1868 bp fragment in lane 8 (I).

The pGT4ΔB candidate seems to be correct: there's one band (E), which has about the same size as the 3758 & 3775 bp fragments in lane 8 (seeD). Apparently the plasmid is only linearised by the BamHI+EcoRI digestion, which must be caused by the absence of the BamHI site..