This paper presents the results of an experimental investigation of the performance of a solar collector outfitted with honeycomb of different arrangements. A complete collector test facility equipped with data acquisition system has been assembled and tested for this purpose. A flat plate solar collector, of 2 m 2 area and 60 mm air gap, has been designed and constructed. The adapted honeycomb unit is structured from polycarbonate sheet of 16 mm thickness containing two rows of honeycomb cells with longitudinal aspect ratio of 0.4. The effect of gap thickness above, below or between the honeycomb units is examined using single or double units of honeycomb. Six arrangements of the single honeycomb unit (SHU), with bottom gaps of 0, 3, 8, 14, 22 and 36 mm, are tested. Six other arrangements of double honeycomb units with different top, middle and bottom gaps are investigated. The experimental data along with correlations obtained by linear regression are presented. The effect of gap thickness for the collector fitted with SHU indicates that the bottom gap is crucial with respect to the heat loss coefficient. The arrangement of single honeycomb with bottom gap of 3 mm is the optimum, as it offers the highest efficiency between all the arrangements and the lowest heat loss coefficient among the other honeycomb arrangements. The results for the collector equipped with two honeycomb units reveal less sensitivity of the heat loss coefficients to variations of the top or middle gaps. Comparing the results obtained for single and double honeycomb units revealed that a compound honeycomb solar collector with air gaps of the right thickness above and below the SHU enhances the collector performance considerably. Also, it affords substantial conservation of honeycomb material because its thickness is about 25% of the total air gap. (C) 2003 Elsevier Ltd. All rights reserved.