all these measures the city was again submerged in 1607, and in a council held by Velasco drainage was agreed upon as indispensable.

The valley of Mexico lies, as is well known, more than seven thousand feet above the sea-level, in a vast basin enclosed by porphyritic ranges, from whose slopes a number of rivers unite to form four groups of lakes, the Chalco-Xochimilco, Tezcuco, Cristóbal, and Zumpango. The first was a fresh-water body, lying two varas higher than the salt Tezcuco, above whose level the last two also rose to the north in their terrace beds four and ten varas respectively. Zumpango received the two largest streams, notably the Quauhtitlan, which contributed a larger volume than that of all the other valley rivers combined. During the rainy season the excess of water overflowed into the Cristóbal lake, which again discharged into the Tezcuco, causing its waters to rise considerably. At certain periods, once in twenty-five years on an average, this overflow proved destructive, especially to the capital, whose main square lay barely four feet above the lake. Taught by experience, the Aztecs had sought to stem the waters with dikes, not only round the city but on the northern lakes. Both of these were, besides, divided into two sections by transverse causeways. Although strengthened and extended under Spanish rule the barriers proved ineffective, as we have seen, and drainage was at last declared to be the only means.

One natural outlet from the valley existed in the small stream of Tequisquiac, but measurements showed that the cost of making it available for drainage would be too great, and that the only practicable point for an outlet was near the village of Huehuetoca, as demonstrated already in 1580 by Licenciado Obregon

i. 728-9. Among the works attributed to him is the presa de Oculma. Inund. de Mex., MS., 356. Cepeda, Rel., 10, is less exact; and so is Humboldt, Essai Pol., i. 209. On the Mexicalzinco branch of the southern causeway two flood-gates were added. Nevertheless this construction proved prejudicial to Xochimilco and adjoining towns. He also began a new aqueduct. Cavo, Tres Siglos, i. 243-4, 264-5; Beltrami, Mexique, ii. 62-3.

CANAL CONSTRUCTION.

9

and Arciniega. By means of a tunnel between the mount Sincoe and Nochistongo hill the ever threatening waters from most elevated northern lakes of Zumpango could be carried through the Tula tributary of Rio Pánuco to the gulf of Mexico. This being decided upon, Enrico Martinez, a Hollander,1o and the Jesuit Juan Sanchez10 submitted plans for the work, one of which embraced also a partial drainage of the middle lakes, while another proposed merely to divert the waters of Rio Quauhtitlan from the Citlaltepec section of Zumpango Lake. The latter was adopted as the speedier and cheaper, and on November 28, 1607, the viceroy broke the first sod in presence of a vast concourse of officials and citizens. The work was intrusted to Martinez,20 who displayed great energy, and set an immense number of Indians to the task, at different points. The expense was covered by a tax of one and a half per cent on the city property, and a levy on wine.21

A canal conducted the waters from the Citlaltepec section of Zampango Lake, or rather from its great tributary, Rio Quauhtitlan, to Huehuetoca, and thence they passed through a tunnel more than a league in length, and four by five varas in height and width, fol

18 Educated in Spain, it seems. He enjoyed the title of royal cosmographer and wrote Repertorio de los Tiempos Historia Natural de Nueva España, Mexico, 1606; Antonio, Bib. Hist. Nova, iii. 564. Humboldt mentions a treatise on trigonometry, Essai Pol., i. 211, but it is probably embraced in the above.

19 Alegre intimates that the plans are due to him, and that he at first had chief control. Hist. Comp. Jesus, i. 438-9. Spanish writers of course prefer to keep the foreigner in the background. Among others connected with the surveys and plans were fathers Mercado and Santos and Doctor Villerino.

20 Sanchez, the associate, soon quarrelled and retired. Torquemada, i. 758. Cavo places the inaugural day on December 28th. Tres Siglos, i. 247. 21 The real estate, valued at 20,267,555 pesos, yielded over 304,000 pesos. Cepeda, Rel., 14. Wine was taxed 50 pesos for every pipe. The clergy were not exempt. Torquemada, i. 758; Recop. de Ind., i. 91-2. The laborers received five reals for seven days, an almud of maize every week, and a pound of meat daily. A hospital was erected for their sick. They came from different provinces, to the number of 471,154, with 1,664 female cooks. Cepeda, Rel., 18. He adds that the actual money paid them between November 1607 and May 1608 was 73,611 pesos. The authoritative writer of Mex., Rel. Estad., 2, declares that 50,000 natives lost their lives during the work, while Cepeda and others maintain that quite an insignificant number perished. They had reasons, however, for hiding disagreeable facts.

lowed by a canal to Rio Tula.22 On May 15, 1608, the first canal was completed, and on September 17th water passed through the tunnel in presence of the viceroy, amidst the rejoicings of the colonists who had reason to be proud of an engineering feat so rare at that time. It was not long, however, before the inefficiency of the work became apparent, the conduit being too small, on too high a level, and so poorly vaulted and faced as frequently to choke with its own debris. The efforts to remedy the latter defect proved of no avail,23 and it was even proposed to construct another channel, for which, in 1611, Alonso de Arias made surveys. Martinez could not well be held to answer, for he had submitted other more thorough plans than the cheap and speedy one adopted. Three years later the celebrated Dutch engineer, Boot, reported in favor of the ancient Aztec dike system for the capital, on the ground that the southern lakes were fully as dangerous as the northern. Martinez agreed to some of his views, but insisted that it was above all necessary to maintain the tunnel outlet.25

22 Cepeda's figures, Rel., pt. i. 25, iii. 21, are 9,600 varas for the tunnel when first opened; afterward reduced by extending it into an open cut. Humboldt is not very exact in giving the tunnel a length of 6,600 mètres, a width of 3.5, and a height of 4.2.

23 Unbaked mud bricks were soon rejected for wood facings, and these for masonry, but instead of an elliptic arch a mere vault was constructed, resting on an insecure foundation, so that the walls were undermined and fell in. The extent of the different facings some years later is given in Cepeda, Rel., iii. 21 et seq.

21 Yet several writers seek to blame him, and assume that the rejected plans had been made by Sanchez. The canal project in 1604 was estimated to require a length of 6 to 9 leagues; now the length of a perfect drainage of the three lakes was placed at 70,000 varas, with a depth of 40. Cepeda, ubi sup.; Gemelli Careri, Giro, vi. 122. By this time the expenditure for the work according to official accounts amounted to 413,324 pesos out of 540,000 collected. In Mex., Rel. Estad., 2, the cost is placed at 1,140,000 pesos up to 1023. Instruc. Virreyes, 262. Gonzalez Dávila makes it 3,952,464 for the first few years, during which 128,630 laborers had been employed. Teatro Ecles.,

i. 2.

23 Martinez prevailed on the authorities to let him perfect the tunnel, but he failed to carry out the agreement, probably because his estimate of 100,000 pesos appeared on closer inspection to be too low. Boot's reports, and the discussion thereon, are given in Cepeda, Rel., pt. ii. 1-17. His views were strengthened by troubles from the southern lake water. Cavo, Tres Siglos, ii. 249. Boot appears to have been retained as active or consulting engineer from 1613 until 1640, with 1,200 ducats pay. Fonseca, Hist. Hacienda, v.

The value of either plan was disputed till Viceroy Gelves, in 1623, caused the tunnel to be closed in order to test the effect of the Quauhtitlan and Pachuca tributaries on Zumpango Lake and consequently on the Tezcuco. The rise proved considerable, and in December came unexpected rains which so increased it that the city was endangered, and the Huehuetoca tunnel had again to be opened;26 and work was renewed upon it in accordance with a neglected royal decree of 1516, although not without much discussion and numerous reports.27

In 1609 occurred a serious revolt among the negroes in the Vera Cruz district. Tired of their masters' yoke, a number of slaves had escaped from different towns and plantations, to unite with their free brethren near the present town of Córdoba, and ensconce themselves among the rugged hills in that vicinity, whence they would pounce upon travellers and settlements. Their leader was an aged man named Yanga, who for thirty years had been seeking to stir his race to united action against the colonists.28 The raids had been endured for some time, attended by the defeat of escorts and improvised troops, under cruel circumstances; but finally the insecurity of the road to Mexico called for stringent measures against the bands, which were growing both in number and daring. Pedro Gonzalez de Herrera of Puebla was commissioned to subdue them, and set forth toward the end of January 1609 with one hundred soldiers, as many volunteers, and a number of native archers, to whom some two

26 The statement of a December flood rests on Gemelli Careri, ubi sup., and has been disputed, but it finds confirmation in the report of a commission of 1624, showing that damage was done to the city by a sudden rise of waters. Cepeda, Rel., pt. ii. 19; Grambila, Tumultos, MS., 11; Ward's Mex., ii. 2527. Early documents bearing on this subject are to be found in Dicc. Univ., ix. 146 et seq.

2 In 1629 came disasters which gave energy to operations, as we shall see. 28 Torquemada, i. 759, intimates that at Mexico also a revolt was projected, for Epiphany, when a king would be elected 'y otros con Titulos de Duques,' etc. It was quickly suppressed.

hundred Spaniards and half-breeds were added from settlements on the way.

A tiresome march brought him near the haunts of the insurgents, though without knowing where or how to meet them. From this dilemma he was relieved by the arrival of a message from Yanga and his military lieutenant Matosa, brought by a captive, who had been defiantly instructed to guide the troops to the foot of the negroes' stronghold, so that they might measure arms with them. Herrera gladly availed himself of this vaunting challenge, to which the chieftain's companions had objected, and in the last week of February he came in sight of the negro camp, on the summit of a mountain. Regardless of the missiles showered upon them, the Spaniards climbed the rugged slope, and though many a one was felled, now by a dart, now by some thundering rock or beam which crushed everything in its path, they persevered and gained the camp, which contained fully three score houses, with church, public edifices, and newly planted fields. The negroes retired to several strong points around, with the loss of quite a number, including several leaders, yet still defiant. Their spirit failed, however, with succeeding reverses, and, as they saw their families falling captive, their houses burned, and their effects seized or destroyed, they submitted terms of capitulation to the viceroy. On condition that Yanga and his free companions be given a site for a new settlement in the neighborhood, they promised to surrender all fugitive negroes in the camps, and thereafter to assist, if duly rewarded, in the capture of any who took refuge in that region. This was agreed to; and soon after they founded the village of San Lorenzo, remaining thenceforth comparatively faithful.29

In the following year a more extensive campaign

29 An alcalde appears to have been appointed from among them, while a neighboring curate attended to their spiritual wants. Alegre, Hist. Comp. Jesus, ii. 10-16.