The two rivers, of which the Rio de la Plata forms the common estuary, differ considerably in their features. The Uruguay has irregular floods, especially in autumn (May) and at the end of the winter (August-October). Low water is in summer (January-February). Its basin belongs to the temperate zone, and does not extend northward as far as the area of tropical summer-rain. The Uruguay also differs from the Parana in its low capacity for transport and alluvial deposit. While the Parana has built up a vast deltaic plain, the Uruguay ends in an ordinary estuary, with rocky or sandy bed and clear water. The estuary of the Uruguay is 130 miles long and five or six miles wide. The eastern shore is rocky and broken. The Argentine shore is low. It is formed in the south by the deposits of the delta of the Parana, while further north, from Gualeguacha to Concepcion, the hills of Entre Rios are hidden behind a screen of flat islands covered with palms, formed by the stuff brought by the streams of Entre Rios. The river-floods are lost in the great sheet of the estuary. The tide in the estuary or a flood in the Parana is enough to turn the current. Maritime navigation goes beyond the estuary and beyond Paysandu, as far as the rapids which prevent further advance at Salto. The twin towns of Concordia (right bank) and Salta (left bank) mark the limit of navigation on the inner course of the river. It begins again above the falls, at Monte Caseros, from which the river-boats go to San Tome and occasionally to Concepcion. Small ships go higher, as far as Salto Grande in Misiones (27 deg. 20' S. lat.).[121] [121] At one time the boats on the upper Uruguay saved transport by going from Salto to Arapehy, midway between Monte Caseros and Concordia (see Isabelle). The navigable system of the Parana is four times as large. The first survey of the river was made about the middle of the nineteenth century by the British Navy. At the beginning of the twentieth century the Argentine Government took up the study of the bed and the peculiarities of the Parana, and the Ministry of Public Works published a map, on the scale 1:100,000, of the course of the river between Posadas and San Pedro, at the beginning of the delta. A precise survey was made, and twenty-six fluvio-metrical scales were established, the zero of which represents mean low-water.[122] Transverse soundings were taken at equal distances of 670 and 1,000 feet, the distance being reduced to 160 and even 80 feet at critical points. Thanks to this work, the Parana is now, no doubt, the best known of all rivers of that size. [122] It is as well to notice that the profile determined by the altitude of the zero of these different scales, or the low-water profile, is of a purely theoretical character. The river is never at low-water over its whole course. The real profile is always varied by slight movements of flood and ebb.