2. The second case is that of a burner consuming 1400 litres (50 cubic feet nearly) of gas, with an illuminating intensity of 14 Carcels (133 candles), placed at the height of 3·20 mètres; this being the class of burner fixed in the Rue du Quatre Septembre. The intensity of light at the foot of the lamp-pillar is 1·367 Carcels (13 candles nearly), and to obtain the light of 0·05 Carcel (0·475 candle) already mentioned as the least intensity enabling one to read, a point must be fixed in a circle of 16 mètres radius from the lamp as a centre. Taking now a group of six lamp-columns, three on each side of the street, and overlapping, as in the Rue du Quatre Septembre, it will be found that the distribution of light is defective. The most brilliantly lighted point at the foot of the column has an intensity of 1·367 Carcels (13 candles), or more than triple that of the darkest point, which has an intensity of 0·5 Carcel (4·75 candles) at 4·58 mètres distance. 3. A lamp of 50-Carcel (475-candle) power, gas or electric, fixed at the height of 8 mètres (26·24 feet). The illuminating intensity at the point vertically under the light is reduced to 0·7 Carcel (6·65 candles); but the light of 0·5 Carcel (4·75 candles) is to be found in a circle of 6 mètres radius from this point. It will therefore be observed that _the distribution of light over the ground is better in proportion as the luminous centre is higher_; but conversely also, _the amount of light thrown on the ground is greater as the luminous centre is lower_. It consequently results that the power of the light and its height should be determined in every case with reference to the effect desired. The method shortly described shows that, in the case of the lighting of the Rue du Quatre Septembre, the mean amount of light per square mètre of the roadway is 855 _décicarcel-cubes_, the best lighted parts having an intensity of 1·62 cubic Carcels, and the darkest portions an intensity of 0·50 cubic Carcel.