and, 4. on the two canals of M. Mier, by which the two lakes of Zumpango and San Christobal may be thrown dry at pleasure.

However, all these multiplied means do not secure the capital against inundations proceeding from the north and north-west. Notwithstanding all the expense which has been laid out, the city will continue exposed to very great risks till a canal shall be immediately opened from the lake of Tezcuco. The waters of this. lake may rise, without those of San Christobal bursting the dike which confines them. The great inundation of Mexico under the reign of Ahuitzotl was solely occasioned by frequent rains*, and the overflowing of the most southern lakes, Chalco and Xochimilco. The water rose to five or six metres† above the level of the streets. In 1763, and the beginning of 1764, the capital was from a similar cause in the greatest danger. Inundated in every quarter it formed an island for several months, without a single drop from the Rio de Guautitlan entering the

* The Indian historians relate, that at this period great masses of water were seen to fall on the declivities of the mountains in the interior of the country, which contained fishes never found but in the rivers of the warm regions (pescados de tierra caliente); a physical phenomenon difficult of explanation, où account of the elevation of the Mexican table-land.

† 16 and 19 feet. Trans.

lake of Tezcuco. This overflow was merely occasioned by small confluents from the east, west, and south. Water was every where seen to spring up, undoubtedly from the hydrostatical pression which it experienced in filtration in the surrounding mountains. On the 6th of September, 1772, there fell* so sudden and abundant a shower in the valley of Mexico, that it had all the appearance of a water spout (manga de agua). Fortunately, however, this phenomenon took place only in the north and north-west part of the valley. The canal of Huehuetoca was then productive of the most beneficial effects, though a great portion of ground between San Christobal, Ecatepec, San Mateo, Santa Ines, and Guautitlan, were inundated to such a degree that many edifices became entire ruins. If this deluge had burst above the basin of the lake of Tezcuco, the capital would have been exposed to the most imminent danger. These circumstances, and several others which we have already adverted to, sufficiently prove how indispensable a duty it becomes for the government to take in hand the draining the lakes which are nearest to the city of Mexico. This necessity is daily increasing, because the bottoms of the basins of Tezcuco and Chalco are continually becoming more elevated from the depositions which they receive.

Informe de Velasquez (manuscript), folio 25.

In fact, while I was at Huehuetoca in the month of January, 1804, the viceroy Iturrigaray gave orders for the construction of the canal of Tezcuco, formerly projected by Martinez, and more recently surveyed by Velasquez. This canal, the estimate of the expense of which amounts to three millions of livres tournois*, is to commence at the north-west extremity of the lake of Tezcuco, in a point situated at a distance of 4593 metrest south 36° east, from the first sluice of the Calzada de San Christobal. It is to pass, first, through the great arid plain containing the insulated mountains of las Cruces de Ecatepec and Chiconautla +, and it will then take

the direction of the farm of Santa Ines towards the canal of Huehuetoca. Its total length to the sluice of Vertideros will be 37,978 Mexican varas, or 31,901 metres§; but what will render the execution of this plan the most expensive, is the necessity of deepening the course of the old desague all the way from Vertideros to beyond the Boveda Real; the first of these two points being 9, 078 above, and the second gm, 181||

* 125,010. sterling. Trans. † 15,067 feet. Trans. The former of these summits, according to the geodesical measurements of M. Velasquez, is 404, and the latter 378 Mexican varas (339 and 317 metres) above the mean level of the lake of Tezcuco.

§ 104,660 feet. Trans.

357.108 inches, and 361.464 inches. Trans.

lower than the mean level of the lake of Tezcuco*. Their distance from one another is al

To complete the description of this great hydraulical undertaking, we shall here insert the principal results of M. Velasquez's survey. These results, on correcting the error of the refraction, and reducing the apparent to the true level, coincide well enough with those obtained by Enrico Martinez and Arias in the commencement of the 17th century; but they prove the erroneousness of the surveys executed in 1764 by Don Yldefonso Yniesta, according to which the draining of the lake of Tezcuco appeared a much more difficult problem to resolve than it is in reality. We shall designate by the points which are more elevated, and by

the points which are less elevated than the mean level of the water of Tezcuco, in 1773 and 1774, or the signal placed near its bank, at the distance of 5475 Mexican varas, south 36° east from the first sluice of the Calzada de San Christobal.

most 10200 metres (33464 feet English). To avoid deepening the bed of the present desague for a still more considerable length, it is proposed to give to the new canal a fall of only 0m, 2 in 1000 metres. The plan of the engineer Martinez was rejected in 1607, purely because it was supposed that a current ought to have a fall of half a metre in the hundred. Alonso de Arias then proved on the authority of Vitruvius (L. VIII. C. 7.), that to convey the water of the lake of Tezcuco into the Rio de Tula a prodigious depth would be requisite for the new canal, and that even at the foot of the cascade near the Hacienda del Salto, the level of its water would be 200 metres* below the river. Martinez could not stand against the power of prejudices and the authority of the ancients! We think that if it is prudent to give little inclination to canals of navigation, it is useful to give in general a good deal to canals of desiccation;

It is to be observed that the vara is divided into 4 palmos, 48 dedos, and 192 granos; that a toise is equal to 3.32258 Mexican varas, and that a Mexican vara is .839169 metres, according to the experiments made on a vara preserved in the Casa del Cabildo of Mexico since the time of king Philip II. Author.

A toise is equivalent to 2.32258 Mexican varas, and not 3.32258. A vara being equal to .839169 of a metre, 2.32258 varas correspond to 1.949 metres 6.394 English feet 1 toise. Trans.

* 656 feet. Trans.