There is an urgent need for methods that can be used to rapidly and nondestructively determine the condition of an old concrete deck beneath an asphaltic concrete wearing course. In recognition of this need, the technique of ground-penetrating radar was investigated. In practice, microwave-frequency impulses of about 1.1 nanosecond pulse width are transmitted into an overlaid bridge deck by a radar transducer that also serves as a receiver. When these electromagnetic pulses are directed through a delaminated concrete area, there is some pulse reflection from the deteriorated concrete, (the more severe the delamination, the more pronounced the reflection), in addition to the normal reflections at the air-asphaltic concrete and asphaltic concrete portland cement concrete interfaces and the reinforcing steel. The reflected pulses are then picked up by the transducer and transformed into the audio frequency range by a time-domain sampling technique and displayed on a facsimile graphic recorder as a pulse reflection profile. Although intended for use on overlaid bridge decks, the technique was experimentally used on three non-overlaid concrete decks and two old concrete deck slabs, in addition to three overlaid decks. To obtain 'ground truths' for comparison, conventional soundings were performed on the non-overlaid decks and slabs and two of the overlaid decks after their overlayments were removed. The results showed that ground-penetrating radar can be used successfully to detect concrete delaminations in both nonoverlaid and overlaid bridge decks, since the delaminations are manifested in the recorded radar pulse reflection profiles as recognizable irregularities in the reflection bands corresponding to the top mat of the reinforcement. These irregularities, or signatures of concrete delaminations, were often in the form of depressions, but in some instances appeared as blurs or breaks in the profiles. It was also found that the radar sometimes missed small delaminated areas of about 1 ft. (0.3 m) width and less. However, this relatively small deficiency does not impair the overall effectiveness of the technique as a nondestructive inspection tool for both types of decks. The experimental procedure can be used as is to inspect decks, if lane closure is not a major concern. However, with little further experimentation, this requirement may be completely eliminated.