The Mokelumne River basin of central California comprises portions of the California Trough and the Sierra Nevada section of the Pacific Mountain system. The California Trough is divisible into four subsections-the Delta tidal plain, the Victor alluvial plain, tlie river flood plains and channels, and the Arroyo Seco dissected pediment. These four subsections comprise the land forms produced by the Mokelumne River and other streams since the Sierra Nevada attained its present height in the Pleistocene epoch. The Victor alluvial plain rises eastward from the Delta plain and abuts on the dissected Arroyo Seco pediment; in the Mokelumne area it is 12 to 16 miles wide and slopes between 5 and 8 feet in a mile. It includes relatively extensive tracts that are intensively cultivated and irrigated with water pumped from wells. The Victor plain has been compounded of overlapping alluvial fans along the western base of the Sierra Nevada. It is prolonged eastward into the pediment by tongues of alluvium along several of the present streams; thus it seems likely that the present stream pattern in the eastern part of the area has been fixed since dissection of the pediment began. Three of the four major streams-the Mokelumne and Cosumnes Rivers and Dry Creek-traverse the Victor plain in trenches which are 15 to 40 feet deep at the heads of their respective alluvial fans but which die out toward the west. The floors of these trenches, the historic flood plains, are from 100 yards to a mile wide. The exceptional major stream, which has not entrenched itself, is the Calaveras River. The Arroyo Seco pediment, which lies east of the Victor plain, was initially at least 8 to 15 miles wide and lay along the western foot of the Sierra Nevada entirely .across the Mokelumne area. Its numerous remnants decline 15 to 35 feet in a mile toward the west. The Sierra Nevada section adjoins and lies east of the California Trough. Its major ridge crests define a volcanic plain whose westward slope is' inferred to have been initially about 90 feet in a mile but is now about 180 feet in a mile, owing to tilting of the Sierra Nevada block in Pleistocene time. In and near the Mokelumne area the Sierra Nevada and California Trough together are roughly coextensive with a single structural unit. The Sierra Nevada constitutes a block that has risen with respect to adjoini;ng valley areas 'by simple rotation or tilting toward the west; it has not been warped or faulted -extensively. It is inferred that this block extends westward beneath the thick alluvial deposits of the trough without material warping or faulting. The oldest rocks of the Mokelumne region are the Carboniferous and Jurassic -rocks that compose the crystalline core of the Sierra Nevada. These are overlain unconformably by sediments of Tertiary age--in upward succession the lone, Valley Springs, Mehrten, and J.Jaguna formations. Of these formations all except the lone are newly discriminated, and type sections are described in the full text. These Tertiary sediments form a great wedge, thinnest along the mountain front to the east, where they have been truncated by erosion. They dip about 2° W. The lone formation (Eocene) consists chiefly of sandstone, clay, and shale; its maximum thickness is 450 feet. The Valley Springs formation (middle? Miocene) overlies the lone formation unconformably. It is composed largely of greenish-gray clay, shale, and sandstone derived from rhyolitic ejectamenta. These rhyolitic deposits are confined to narrow channels in the higher part of the Sierra Nevada, but they spread fanlike over the lower western edge of the mountain block, where they attain a maximum thickness of 525 feet. The Mehrten formation (upper? Miocene and lower Pliocene?) comprises the andesitic rocks that constructed the Sierran volcanic plain. In the Mokelumne area it consists chiefly of sandstone and siltstone but includes, as a minor though conspicuous part of the formation, layers and tongues of resistant breccia or agglomerate, which are presumed to have originated as mud flows. Nonfragmental andesite is not known to occur in the Mokelumne area, although several possible vents occur farther east. In the eastern part of the area the Mehrten formation truncates in turn the Valley Springs and lone formations and the pre-Cretaceous rocks; in the western part the Mehrten formation (andesitic) interfingers with the underlying Valley Springs formation (rhyolitic). Its maximum measured thickness is 400 feet. Few of the irrigation wells are so deep that they can be said with assurance to reach the Mehrten formation. The Laguna formation (Pliocene? and possibly lower Pleistocene) comprises poorly sorted, nonandesitic fluviatile sedimentary that overlie the Mehrten formation. It is inferred to be essentially parallel to and tilted equally with the Mehrten formation and to be about 400 feet thick. The Arroyo Seco gravel (presumably middle Pleistocene) veneers the Arroyo Seco pediment. At its easternmost outcrops the formation is composed of pebbles, cobbles, and boulders in a matrix of brick-red sand and silt; farther west, down the slope of the pediment, it becomes pr9gressively finer. It is inferred that the Arroyo Seco gravel is a coarse fraction of the rock waste that was transported from the Sierra Nevada after the Sierran.block was tilted in Pleistocene time. It is inferred further that the correlative of the Arroyo Seco gravel in the California Trough is a wedge-shaped mass of sediments whose base is the tilted Laguna formation and whose top can be interpolated by projecting a hypothetical surface through the remnants of the pediment. The Victor formation comprises the fluviatile sand, silt, and gravel that built the Victor alluvial plain over the hypothetical equivalent of the Arroyo' Seco gravel along the axis of the California Trough and against the western front of the dissected pediment to the east. The formation is thought to be about 100 feet thick along the western margin of the Mokelumne area, according to an estimate based upon projecting the slope of the Arroyo Seco pediment westward beneath the Victor plain. The Mokelumne area lies on the fertile central plain along the Mokelumne River about the city of Lodi, in northern San Joaquin County, and has been intensively developed for the cultivation of grapes, deciduous fruits, and other crops. Of necessity its great productiveness is maintained by irrigation. Extensive irrigation from wells began about 1907 and has increased steadily until in 1932 about 50,000 acres (80 percent of the area) was watered in that manner. The specific question at issue is the extent to which the supply of ground water and hence the productiveness of the area are dependent upon the water flowing in the Mokelumne River and the extent to which that productiveness may be influenced by regulation of the stream--:in particular, by the substantial regulation of the river that is accomplished by the Pardee Dam of the East Bay Municipal Utility District, which began to function in March 1929. The depth of 1,447 irrigation wells in five townships in the central part of the area (T. 3 N., Rs. 6 and 7 E., and T. 4 N., Rs. 6 to 8 E.) ranges from 20 to 910 feet. About half the wells bottom within a 100-foot zone whose base is 75 feet below the projected Arroyo Seco pediment; essentially that zone constitutes the Victor formation. Only 6 percent of the wells bottom within the next lower 25- foot zone, but the percentage increases sharply for depths still greater; it is inferred that impervious strata are relatively persistent between 75 and 100 feet below the projected pediment and that these are the uppermost part of the Arroyo Seco gravel. Of 580 observation wells known to bottom in the Victor formation, essentially all appear to indicate a regional water-table stage; thus the water is essentially unconfined. On the other hand, nearly all wells so deep that they reach the Arroyo Seco gravel or some underlying formation tap confined water. Near the Mokelumne River the water levels in these deep wells stand below the water table, which is semiperched. In most deep wells remote from the river the water level stands above the water table except during the pumping season. Fluctuations of ground-water levels are ascribed to moving or changing load on the land surface, earthquakes, variation of barometic pressure, ground-water draft by vegetation, infiltration of rain and certain indirect effects of rainfall, infiltration of water applied to the land for irrigation, variation in the discharge of streams, and pumping from wells. In the eastern part of the central district, between Clements and the vicinity of Lockeford, it is inferred that (1) the river and the water in the alluvium of the flood plain are not insulated from the water in the sediments that form the adjacent Victor plain; (2) locally if not generally, however, there are discontinuities in pervious strata along the outer margin of the flood plain, where the water table passes from the alluvium into the enclosing sediments, so that percolation of ground water is impeded materially at that margin; (3) rising river stages set up ground-water waves that store relatively large volumes of water in the alluvium close to the river, whereas falling stages cause much of that stored water to percolate back into the river, weeks and even months lapsing before the ground-water stage becomes steady within the flood plain; and (4) seepage loss from the river into the alluvium tends to be intermittent and to alternate with seepage gain, the rate of loss or gain lagging weeks or months behind the fluctuations of river stage and lagging more for moderate changes at low stage. However, in the succeeding reach downstream as far as Woodbridge, it is inferred that percolation of ground water is not impeded generally along the outer margin of the flood plain and that the river tends to lose almost continuously by seepage rather than intermittently, although the rate of loss fluctuates somewhat in response to changing river stage. The yearly pumpage for irrigation has been as much as 114,600 acre-feet (1928- 29), and there have been as many as 2,500 wells equipped with irrigation pumping plants (1931). Commonly the wells are pumped only in daylight and are idle over week-ends and holidays, also during and after protracted rainstorms in the early part of the season. In a small district near Victor pumping in recent years has begun in January or February, has reached its height in March, and largely has passed by April. In outlying districts general pumping has begun as late as May, reached its height in June or July, and waned by September. Since 1907 the water table appears to have declined steadily in most of the Mokelumne area except along the river. The decline was least in the Woodbridge Irrigation District, where in four typical wells. the average decline from 1907 to 1937 was 3 feet, or 0.15 foot a year. Among 18 shallow wells in the district of most intensive pumping the average recession of the water table from 1907 to 1927 was 11 feet, or 0.55 foot a year; the greatest measured recession was 15 feet, or 0.75 foot a year. From 1927 to 1933 the water table declined 5 feet or more over most of the central pumping district except within 2 miles of the Mokelumne River, and the greatest measured decline was 9 feet. The area of material recession ,extends 4 to 7 miles eastward beyond the central pumping district, whence it is inferred that pumping has drawn gradually on remote ground-water storage. It is inferred that the Mokelumne River ordinarily has been a losing stream between the Mehrten dam site, near Clements, and the Woodbridge Dam, the area that received the percolate having been triangular with its upstream and having included about 5,200 acres of the flood plain and 36,500 acres in outlying districts to the north and to the south. Mean fluctuations of the water table within the area receiving percolate from ihe river are believed to indicate that relatively little water is drawn from outside the area. Accordingly, simple storage methods are competent for a ground-water inventory. It is inferred that the rate of seepage loss from the river depends jointly upon river discharge, stage in the Woodbridge Reservoir, and groundwater pumpage. The foregoing inferences lead to the following conclusions with respect to ground-water replenishment by seepage loss from the river in the intensively cultivated district about Lodi: (1) The annual replenishment has tended to increase for at least two decades, owing to the gradual increase in head between surface water and ground water as ground-water levels have been lowered progressively by pumping; (2) annual replenishment has tended to increase, especially in recent years, owing to gradually prolonged use of the Woodbridge Reservoir, for thereby a relatively large wetted area and great differential head have been maintained for an increasing term; (3) the rate of replenishment tends to be greater under regulation than under the so-called natural regimen, to the extent that regulation has maintained a moderately large wetted area and stage in the river through the later part ·of each pumping season, whi1e the ground-water levels have been lowest. Moreover, for any particular yearly run-off below the Mehrten dam site, the replenishment by seepage would tend to be greater under the regulated regimen to the extent that fluctuations in discharge were suppressed, for the greatest yearly mean stage and mean wetted area would be afforded by constant discharge. -Thus, diverting water out of the Mokelumne River Basin at the Pardee Dam does not necessari1y-entail a diminution in ground-water replenishment by seepage loss along the lower reach of the stream, at least in the replenishment beneath the Victor plain above the gaging station at Woodbridge. Rather. the Pardee Dam affords a means for so regulating the discharge as to effect a maximum ground-water replenishment with-a given run-off in the natural channel. Bodies of ground water perched above the regional water table are common in the Laguna formation, especially in its lower part. Conspicuous bodies occur about 3 miles south of Clay, in a district between 1 mile and 5 miles south of Clements, and along Dry Creek in T. 5 N., Rs. 7 and 8 E. From the relation between the water table and the piezometric surface for water confined in deep aquifers, the area receiving percolate from the Mokelumne River may be divided roughly into (1) a central area, extending not :p1ore than half a mi1e beyond the flood plain, in which the piezometric surface is inferred to have stood below the water table throughout the term of the investigation and hence in which the difference in head has favored the percolation of water from shallow beds into deep beds in all seasons, and (2) an outlying area in which the difference in head likewise favors downward percolation into deep beds during the pumping season but favors upward percolation during the nonpumping season. This outlying area includes about 75 percent of the segment of the Victor plain that receives percolate from the river. From 1927 to 1933 the subartesian head that existed during the nonpumping season in the area remote from the river tended to increase; it is therefore inferred that the relative opportunity for seasonal recharge of the shallow water-bearing beds by underfeeding has likewise tended to increase. On the other hand, the negative differential head in wells near the river also has tended to increase; thus in this central area the opportunity for discharge of water from shallow beds by downward percolation has probably tended to increase. It is believed that ground-water storage within the area near the river is not decreased materially by" discharge westward through deep pervious beds, also that the yearly addition to ground-water storage in the outlying area by deep percolation from a remote easterly source is scant and for all practical purposes is offset by downward percolation along the river.