Artesian Wells of the Christchurch Area.

Transactions and Proceedings of the Royal Society of New Zealand, Volume 56, 1926, Page 369

Artesian Wells of the Christchurch Area. By F. W. Hilgendorf, M.A., D.Sc., F.N.Z.Inst. [Read before the Canterbury Philosophical Institute, 3rd December, 1924; received by Editor, 31st December, 1924; issued separately, 31st March, 1926] Contents. A. Objects and Methods. B. Water-level Observations. 1. Rain Effect. 2. River Effect. 3. Barometer Effect. 4. Industrial Effect. 5. Evaporation Effect. 6. Tide Effect. B. Water-level Observations—continued. 7. Pressure Effect. 8. Earthquake Effect. 9. Correspondence between Rises in the River Avon and those of the Wells near by. C. The Chemical Observations. D. Summary and Conclusions. A. Objects and Methods. The investigations here recorded were carried on by the writer with the assistance and advice of the Artesian Wells Committee of the Canterbury Philosophical Institute, the monetary expense involved being met by a grant to that committee from the New Zealand Institute. The work dealt with in this report was started in 1921, and was carried on almost continuously till November, 1924, by which time it appeared that the number of new facts likely to come to light by the method used was not commensurate with the labour involved in discovering them. The writer has been making observations on the Christchurch artesians for the past fourteen years, and the information gained has been largely drawn on for comparison with the results of this series of observations. The ultimate aim of the research was to enable a prophecy to be made as to the future fate of the wells—whether the supply is likely to become insufficient as the draft on it increases, and whether the level to which the water rises is likely to diminish in the course of years. For this purpose it was sought to find the source of the water supplying the wells, and the quest was started along two lines: first, recorders giving continuous readings were attached to a number of wells, and meteorological observations were made or collected, to see if there were any correlation between the level of the waters in the wells on the one hand and such things as rain, evaporation, river-floods, &c., on the other; and, second, chemical analyses were made of the waters from various wells and the possible sources of supply to see if there were any chemical similarities between their compositions. B. The Water-Level Observations. These were made on eight wells varying in depth from 60 ft. to 344 ft., and situated either in Christchurch or at Lincoln, twelve miles from Christchurch. Wells have to be out of use for the purposes of water-supply while continuous records are being taken of their static head, and this largely restricts the number of wells available, so that frequently one has to go far afield to secure wells of the desired depth. The arrangement adopted for securing the records consisted of (1) a float-chamber at the top of the well-pipe: (2) a float connected by cords and pulleys with a pencil, to which movements of the float imparted a

series of vertical movements: (3) a drum rotating on a vertical axis, driven by clockwork, and carrying a paper on which the pencil marked the records. The float-chamber was 6 in. in diameter and 6 ft. in height; it was so arranged that at the time of its erection the static level of the water was about the middle of its height, so that the water might rise or fall about 3 ft. before putting the float out of action. The float weighed 3 lb. and was connected with the pencil by thin steel wire, which was proved, by trial of the apparatus on standing water, to impart no motion to the pencil by such alterations to its length as were produced by the weather variations to which it was subjected. The rotating drums were 18 in. in height and about 1 ft. in diameter: they rotated once a week, were driven by hanging weights, and their timing was regulated by their being belt-connected to large kitchen clocks of the “Big Ben” type. These recorders worked perfectly well indoors, but in the temporary shelters erected over the wells they were apt to stop working in wet or cold weather, so that important records were sometimes lost, and the necessity for winding them daily made them expensive to look after when they were far apart. In the end a barograph, in which the vacuum-box was disconnected from the recording-pen, was modified for use on the wells, and its weekly winding and regular running made it perfectly satisfactory in action. The instruments (three in number) were first attached to three wells in Christchurch, all situated near Canterbury College; but many small movements, and the beginnings of larger ones, are, in the town wells, masked by the effects of intermittent pumping from neighbouring wells. For the study of these smaller fluctuations the instruments were moved to Lincoln, where pumping from the wells is not practised; and this removal was rendered additionally desirable because there were indications of fundamental differences between the wells at Christchurch and those at Lincoln. In the following paragraphs the behaviour of the town wells and those at Lincoln are frequently contrasted. At one stage of the observations a recorder was placed on the River Avon, and some of the observations made are recorded in the final paragraph of this section. 1. The Rain Effect. (a.) The Wells at Lincoln. Size of the Effect.—That rain raises the level of the water in the wells is common knowledge. The amount of rise produced by 1 in. of rain varies greatly with the condition of the soil and subsoil before the rain concerned. If the rain comes after a drought its effect on the well is only slight; if it comes after previous rains its effect is much more pronounced. This is probably because in the first case all the rain does not reach the free subsoil water at the outcrop of stratum whose water supplies the well, but is used up in moistening the soil and subsoil: if this is already wet by previous rains nearly all the subsequent fall reaches the water-table. In many places a saturated soil and subsoil would increase the surface run-off instead of raising the water-table, but over the greater part of the plains above Lincoln the soil is so gravelly and pervious that there is no run-off at all. Thus of two equal rainfalls occurring at a short interval of time the second has an effect much greater than the first, because the second fall is not called upon to moisten the soil before it reaches the water-table. It is therefore not possible to state

Fig. 1.—Fluctuations of various wells.

the effect of any given amount of rain on any given well. Consecutive rains have a cumulative effect. This cannot be made clearer than in the graphs illustrating former papers on this subject (2, 3) or that given in less detail in fig. 1 of the present article. The maximum variation in well-level recorded in these observations is in a 344 ft. well at Lincoln College, which stood at 8 ft. 2 in. above zero-mark in 1912, and fell to 9 in. below zero in 1923. This must not, however, be considered a permanent fall, because the great height of 1912 was attained only after two annual rains of 32 in.—28 per cent. above the normal 26 in. In 1911 the water-level was only 2 fft. 6 in., and two wet seasons produced a rise of nearly 6 ft. It is almost certain that similar seasons in future would produce a similar rise, and that much of the uneasiness concerning the present level of the wells will be allayed when a few wet years have again been experienced. The year 1914 had only 20 in. of rain, and 1915 only 14 in., so that the wells fell seriously. Since then we have had almost normal rainfalls, but nothing to counterbalance the serious drop in the wells following on the years of drought. The maximum rise after a given rain was recorded on a 308 ft. well at Lincoln, where a rise of over 18 in. followed the 4 in. rain in May, 1923. How much more than 18 in. the rise was could not be ascertained, because the water overflowed the top of the pipe (which was already 25 ft. above the ground) and it became necessary to shut off the well. This particular well (the old creamery-well at Schaeffer's Corner) was sunk in 1902; its depth was 308 ft. The water rose 25 ft. above ground-level, and the 2 in. pipe discharged 25 gallons per minute at the surface. The height is still 25 ft. after a rain, but the surface discharge is only about ½ gallon per minute. This is a case of blockage of the pipe, and the flow could be restored by reboring the present pipe. The reduced flow of a well whose static head is also reduced cannot be restored by any manipulation so far known. At the time of the 18 in. well-rise described the rainfall at Lincoln was only 4 in., but it was 13 in. at Darfield, which may be considered as nearer the outcrop of the stratum tapped by the well. The College well (344 ft.) rose about 17 in. on the same occasion. There also the exact rise could not be measured, as the water had sunk to below zero on the gauge before the rain mentioned. The amount of rise per inch of rain is governed by the distance the water runs over impervious strata before it falls into the pervious one that is tapped by the well. The rise of any well therefore varies with the nature of the layers at the outcrop of the stratum tapped, but all near-by wells sunk to the same depth probably react similarly. Hutton (4) notices that his deep well fluctuated more than his shallow one, and I have found that in general deep wells rise more than shallow ones per inch of rain. While the greater number of shallow wells in town is a sufficient explanation of the phenomenon there, other factors must be called in to explain it in the country, where shallow wells are not in excess. It may be supposed that the strata tapped by the shallow wells reach the surface near the artesian area, while those tapped by the deep ones reach the surface higher up the plains. The farther up the plain (within limits), the more shingly is its surface; the nearer the sea, the more is its surface clothed with soil. Rain falling close to Lincoln sinks slowly and imperfectly to the strata tapped by the shallow wells: rain falling farther away sinks rapidly and completely to those tapped by the deeper ones.

It is probable that the rain feeding the wells falls within, say, ten miles of Lincoln or Christchurch, and not far up the plains, or in the mountains, as is sometimes supposed. A popular belief connects the snows on Mount Torlesse with the artesian water-supply, but this is almost certainly fallacious. All the snow that melts off Mount Torlesse must be caught by the Kowhai or the Waimakariri, and its effect on the wells must be by way of the river. To the river, however, the effect of Torlesse snows is negligible in comparison with the snows it receives from other quarters, and so the effect of Torlesse snows on the wells must be negligible too. Benmore snow may conceivably affect the Lincoln wells, because, although much of it finds its way to the Kowhai and Waimakariri, a certain amount is fed to the Selwyn, which loses itself in the middle plains, and so may affect the seaboard wells. Years of observation have, however, failed to show any relation between the melting of the snows on the frontal range and the fluctuations in the water-level of the artesians. It is true that the Lincoln wells often reach their annual maximum in October-November, at the end of the melting of the mountain snows, but an examination of fig. 1 will show that there was a continuous fall during 1914 and 1915 with no maximum in early summer, although snow-melting took place as usual; and the same is true of 1923. The graph, moreover, shows the all-sufficiency of the rainfall near Lincoln to explain all the major fluctuations of the well-level, without calling in the aid of any such supposititious factors as the snow on the distant ranges. Time of the Rain Effect.—Next to the amount of the effect due to rain, the length of time taken for the influence of the rain to be felt on the well calls for consideration. To find out this time is not as easy as was anticipated when the recorders were installed, for the wells respond to barometric changes as well as to rainfall. The wells fall when the barometer rises; and the barometer almost invariably rises when rain commences. Thus the first effect of rain is to make the well-level recede, and this effect has to be overcome before the rain rise is observed. However the barometric effect on certain wells has been determined and can be allowed for with fair accuracy. In the best case noted the fall due to barometric rise was checked, stopped, and reversed within two hours of the start of a fall of rain; but in other cases the effect was much more retarded. When the ground was very dry in May, 1923, and then 4 in. of rain fell in one week, it took two days before any rise was recorded—and this on the same well that on another occasion had responded to rain within two hours. It would appear that the time taken to cause a rise depends entirely on the time taken for the rain to reach the water-table where the water-bearing stratum outcrops, and that the effect is then immediately imparted to all parts of that water-body, including the part standing in the well-pipe. This would require that we suppose the water-supply to be practically an underground lake among the stones of the gravel-beds; that water is pouring into one-end where the bed outcrops on the plain, and out of the other end by means of springs and flowing wells. The spaces among the stones must be very great to allow of such free movement of the water as this rapidity of effect implies, and that large spaces do exist is emphasized by Dr. Chilton's record (1) of subterranean Isopods up to nearly 1 in. in length from the gravels tapped by the shallower wells. (b.) The Town Wells. Size of Rain Effect.—This is much less in town wells than in those at Lincoln, doubtless because many wells are sunk to the same stratum, and

when a rise takes place every well flows more freely and so any one of them is prevented from rising remarkably. No records were being taken on town wells in May, 1923, when a rise of over 18 in. was recorded in a country well, but inquiries showed that the rise was probably about 6 in. Speight (5) notes “a jump of 6 in. after several days' rain,” and this must be considered a large rise for a town well. The great steadiness of the town wells as compared with the country ones is well shown in fig. 1. This is of great interest and importance, for it probably shows that the town wells are normally at their basic level, below which it is improbable any influences will lower them. The source of supply is of such constancy that periods of drought affect them but little. The country wells, being few in number, are normally above their basic level, and therefore are quickly lowered by rainless seasons. This was probably the condition of the town wells in their early days. Time of Rain Effect.—The town wells are so much influenced by pumping from neighbouring wells to below their static head that all minor fluctuations caused by the weather, and the beginning of all larger ones, are totally obscured. There is therefore no evidence of the time after rain that the town wells begin to rise. Symes's (6) suburban well was partly freed from the effect of near-by pumping, and he got the record of a rise within one hour of the commencement of rain—which is of the same order as the two hours noted at Lincoln. 2. The River Effect. (a.) The Wells at Lincoln. These might be affected by either the Waimakariri or Selwyn. As the Waimakariri leaves the mountains it is pointing straight towards Lincoln, and does not swing off to the north till it reaches Halkett. From Halkett to Lincoln is about fifteen miles, and the fall of the surface of the plain, and probably of the underlying strata as well, is direct from Halkett to Lincoln, and not from Halkett to Kaiapoi, as the river now runs. It is therefore quite conceivable that seepage from the river reaches Lincoln. The Selwyn runs underground for the middle fifteen miles of its course, disappearing under the shingle above Greendale, and reappearing above Ellesmere. There is every probability that this underground water seeps away to be tapped by wells, and does not all rise again in the river-bed. Mr. W. Turner, of Selwyn Railway-station, kept a record of the floods in the Selwyn for me for some years while I tried to find a reaction to them in the Lincoln wells. No effect from Waimakariri or Selwyn has ever been recorded as the result of direct observation in any well at Lincoln, and the five months graph shown in Trans. N.Z. Inst., vol. 44, p. 147, is pretty conclusive evidence that no such effect occurs as far as the larger river is concerned. In 1914 we had 20 in. of rain, and in 1915 14 in., compared with our average of 25 in. As a result of this, some springs at Brookside, two miles from the Selwyn River, went dry. In 1916 a heavy rain occurred, causing a flood in the Selwyn, and a fortnight later the Brookside springs started running. This is sufficient evidence that the Selwyn feeds underground reservoirs; but if the flood-water when underground travels at the rate of only a mile a week, it would be impossible to trace cause and effect in the case of the Lincoln wells, which are twelve or fifteen miles away from any possible seepage-bed in the Selwyn.

The slowness of travel of the water on this occasion serves to emphasize the rapidity of the rain effect, and to support the idea that the ordinary source of supply is either a virtual underground lake, or freely-moving underground streams whose intake is quite close to the wells. In the present case one may suppose seepage from the river to these springs to take place only from flood-waters, and therefore through usually dry beds which would greatly retard the water's flow. While floods in the rivers cannot be traced in the fluctuations of the wells, yet a steady seepage is to be regarded as very probable. From early in 1914 till February, 1916, the 334 ft. well at Lincoln College fell steadily and uninterruptedly from 4 ft. above the zero-mark to 4 in. below it. Then, though the weather continued dry, and although a continuation of the fall might have been expected, the diminution of static head ceased, and the level remained constant until the 28th May, when heavy rains caused the well to rise. In November, 1922, the well-level had again fallen below the arbitrary zero, and again remained about 4 in. below until late in May, 1923, when the heavy rains of that month caused the water to reappear in the gauge. On neither of these occasions did a ram whose intake-pipe is only 11 in. below zero on the gauge cease to work. That on two occasions of long and constant fall, the fall should be arrested at the same level without any rain to check it, clearly suggests a source of supply more constant than the rainfall alone. It looks as if there were a constant supply held by river-seepage to a certain level, and when heavy rain falls this level is raised considerably, because the few wells at Lincoln (four or five per square mile) do not let the water away quickly enough to check the rise. In course of time, however, the increased level due to the rain is reduced, and the wells then stand, and continue to stand, at the level of the water as supplied by the rivers. This evidence of the river-supply would not be sufficient to prove it if Lincoln wells alone were considered, but the evidence from the town wells to be next considered is very strongly in this direction. (b.) The Town Wells. Here there are many thousands of wells per square mile, and their combined flow is much over 10,000,000 gallons per day. These wells are very constant in level, as shown in fig. 1. They rise with rain, but to a much less extent than the country wells, and quickly fall back again to their constant level, below which their continuous and voluminous flow is not able to reduce them. The evidence in favour of river-supply is practically conclusive. The rainfall is, indeed, sufficient to supply the flow, but the rainfall is intermittent, while the wells, once the rain has ceased for a few days, remain at a level that is practically constant. The relative constancy of the town wells compared with the country one (fig. 1) indicates that the former are using chiefly river-water and the latter chiefly rain-water. No observation of a well's rising after a Waimakariri flood has yet been made in any well farther from the river than Belfast, despite the general impression that such a rise takes place. Speight (5) suggests that barometric influence may cause the apparent connection, and the next section will show that this surmise is probably correct. 3. The Barometric Effect. (a.) The Country Wells. The effect of the variations of the barometer previously recorded (1) were found again in every record made during the present series of

observations, but were differently developed in different wells. The Lincoln College well (344 ft.) rises four times as much as the barometer falls, but a near-by well of 60 ft. in depth showed a water-rise only equal to the barometer-fall. A typical week's graph for this well is reproduced (fig. 2), the well reading being inverted. (b.) The Town Wells. Here no connection between barometer and well-level has ever been recorded, because the effect of pumping adjacent wells below their static head masks all small fluctuations to such an extent as to make accurate measurements impossible. The barometric effect must, however, be present in full force, and if it has anything like the effect it has at Lincoln the result must be quite striking. Hutton's graphs (4) show that rises of 4 in. are quite uncommon, and would cause a noticeable difference in the flow of a well or the working of a hydraulic ram. But a rise of this sort, if not of this amount, must be produced every time the barometer falls 1 in., which it does every time a north-wester blows; and this is doubtless the cause of the general observation that rams work better during a north wester or when the Waimakariri is in flood. Fig. 2.—B, barogram; W, inverted record of 80 ft. well at Greenpark. 4. The Industrial Effect. This is the name suggested by Dr. C. C. Farr for the interference with the static head of the well under observation by pumping other wells to below their static head. For the circumstances concerned it is a better term than “interference,” which includes also the effect caused by the sinking of new wells and the opening and closing of old ones in the neighbourhood of that under observation. (a.) The Country Wells. No industrial effect was observed here, because pumping direct from wells is not practised in the neighbourhood. This absence of industrial effect allows other effects to be more closely seen.

(b.) The Town Wells. The industrial effect was first noted by Captain Hutton (4), who states that the shallow well at the Christchurch Museum rose every evening, and who was able to assign this to industrial causes, because no corresponding fall took place on Sundays or holidays. Hutton's “evening rise” was perfectly well marked, and it was to see if the rise were really an industrial effect that observations on country wells were first started. There appeared some slight evening rise in my first observations (2), but the present series shows it to be much greater in town, and to be really industrial in origin. Instead of an “evening rise” it should be called a “morning fall,” because the higher level is the normal one, and the lowering is produced by pumping. The Museum wells, observed by Captain Hutton and Professor Speight, and the three wells recently under observation. are all within half a mile of the industrial quarter of the town, and so they promptly feel the effect of pumping. The places and times of pumping have been traced, and the accord between start of pumping and falling of well is almost exact. Symes's (6) records show daily curves with maximum and minimum similar to the wells observed nearer town, but owing to his well being about three miles from the industrial part of the town the curves are somewhat more gradual. A week's graphs (not the same week) from Symes's Papanui well, and one from a well near the Hospital, are here produced for comparison (fig. 3). The absence of a fall on Sunday is again to be noted. Fig. 3.—Industrial effect on—A, well near Hospital; B, Symes's Papanui well. These fluctuations will now be seen to mask any due to barometric effect. 5. Evaporation Effect. It was once suggested that the fluctuations of a well's level would be completely understood if one knew the rainfall and the evaporation. This was an attempt to explain the fact that rainfall in dry weather has a much less effect on the wells than the same rainfall in wet weather. An

evaporimeter was therefore installed, consisting of a tank 4 ft. square and 2 ft. deep, placed in the middle of an open field, and kept filled with water to within 2 in. of the top. Observations were made during five years, and the records show for each month the total evaporation, the rainfall, and the “effective rainfall”—that is, the rainfall minus the evaporation. When the well-levels were graphed against the “effective rainfall” there was no appearance of accord. Average Monthly Rainfall and Evaporation from a Water-surface at Lincoln, N.Z., for the Years 1916–23 inclusive. (All in inches.) Month. Average Rainfall. Average Evaporation. Average Effective Rainfall. January 3.4 5.4 −2.0 February 1.6 4.5 −2.9 March 1.0 4.0 −3.0 April 1.8 2.8 −1.0 May 2.3 2.1 + 0.2 June 2.2 1.1 + 1.1 July 1.8 0.8 + 1.0 August 1.6 0.8 + 0.8 September 2.5 2.7 −0.2 October 1.6 3.5 −1.9 November 2.1 4.1 −2.0 December 1.5 5.5 −4.0 It is thus seen that the effective rainfall is highest in midwinter, while fig. 1 shows that the wells are highest in early summer. The fact is that it is not the evaporation in the month that the rain falls in that modifies the action of the rain upon the well, but the evaporation in previous months. If this has been slight, a normal rainfall has a great effect. Thus the lack of evaporation during winter causes the spring rains to influence the wells so that their maximum height is reached in October and November: the great evaporation during summer causes the autumn rains, though equal to the spring ones in amount, to have little effect, so that the wells reach their minimum in April and May. The variations in height of the wells are governed by the rain modified by the previous evaporation. The utility of a rainfall from an agricultural point of view is inversely proportional to its effect on the well-levels. The average rainfall for the five years concerned was 23.4 in., and the average evaporation 37.3 in. 6 Tide Effect. Professer Speight shows (5) that a tidal effect amounting to 18 in. rise at high tide, and extending to three miles from the coast, occurs at New Brighton. The wells noted in the present series of observations are too far from the sea to show any tide effect.

7. Pressure Effect. Symes (6) shows that a tram-car, especially the heavy water-carrying type, produced a distinct wave-like movement in the water in his well, 30 ft. from the tram-line. Speight also refers to trials in Japan showing rise of water in response to artificial loading. This might account for some of the rise in coastal wells due to the rising tide, but the saltness of the New Brighton wells is against the pure pressure explanation. The rise of wells in response to rain may also be partly due to the weight of water falling on the earth—partly only, because the rise in the wells goes on long after the rain has ceased falling. The first rise, however—that occurring within the first hour or two of the rain—might conceivably be associated with pressure effect. I was at first inclined to think that, if the well-rise were due to pressure, every rain should produce the same rise; but, again, it is conceivable that air in the soil, after a period of drought, might cushion the rain's pressure so as to postpone its action. Pressure may therefore have some slight influence in causing the first rise due to rain; but the magnitude of the total rise, its continuance after rain ceases, and the want of any sign of break in its continuity, all indicate that the pressure effect is of very little significance. 8 Earthquake Effect. Mr. L. P. Symes has kindly allowed me to insert here a note on an observation he made in 1917. In August of that year he was still running a continuous recorder on his Papanui well, as he has previously described (6). At about 0.45 on Monday, the 6th of that month, the pencil attached to the float on his well suddenly gave a rise and fall of ⅔ in., and continued to oscillate for twenty minutes. After the first great throw of ⅔ in. the next vibrations were of ⅛ in., and they gradually diminished until they disappeared. Symes carefully checked the time of this disturbance in his well and found it to synchronize exactly with an earthquake recorded by the Christchurch Milne seismograph, so that there is little doubt that the disturbance he noted was in reality caused by an earthquake 9. Correspondence between Rises in the River Avon and those of the Wells near by. It was not expected that fluctuations in the River Avon would have any effect on the water-level of the wells, but it was thought that the variations of level of the one might elucidate those of the other, and so a recorder was placed in position on the banks of the river, just above the Hospital, and continuous readings were taken for some months. No relation between river and wells was observed. The river-level remains stationary for days on end, drawing a straight line on the graph, strikingly different from the agitated curves produced by the fluctuations in the adjacent wells. The only fluctuations observed are—(a) Those due to rain. The rise commences within half an hour or less of the beginning of a storm, and is in the neighbourhood of 7 in. per inch of rain if this falls within twelve hours. A rise of 7 in. in two hours has been recorded, but this was a torrential downpour of 0.7 in. (b) Those due to the river being cleared of weeds some distance below the recorder. On the 18th October, 1921, the river fell 4 in. in twenty-four hours, and

remained permanently lowered for some weeks. On that day weeds were being cut at Barbadoes Street over a mile below the recorder; and, although it hardly seemed likely that the influence of the cleaning would be felt so far up, no other explanation of a permanent depression of level suggests itself. (c.) Those caused by manipulation of the dam towards the headwaters of the stream. This sometimes causes the river at the Hospital to fall 2 in. or 3 in. in the course of an hour, and to recover its original level in the course of another hour. A sudden rise of about 2 in. is explained by the same cause. These changes usually take place at Saturday and Sunday midnights, and have no permanent effect on the river-level. Fig. 4.—Fluctuations in level of River Avon: rain on Wednesday, dam-manipulation on Friday. A fragment of a graph showing the rise due to 0.3 in. of rain, and also a combined fall, recovery, rise, and recession, due to dam-manipulation, is given in fig. 4. C. The Chemical Observations. It was thought that chemical analyses might show the source of supply of the Christchurch wells, and so samples were taken of water from the Waimakariri River; from a shallow non-flowing well at Harewood, about two miles from that river; from a 400 ft. well in Fletcher-Humphries' yard in Cathedral Square; and from the 344 ft. well at Lincoln College. The actual collection and analyses were done by Mr. M. J. Scott, of Lincoln College, for whose great care in the matter I am deeply grateful. Many of the individual trials were done in quadruplicate, and none in less than duplicate. There is no suggestion in the figures that the Christchurch waters have a different source of origin from the Lincoln waters. The solids are obviously taken up during the passage of the water through the gravels of the plains, as the deep wells contain twice as much solid matter in solution as the river-water does. The figures will be seen to show a gradual increase in total solids the farther the well is from the river, and in general a similar gradual increase in each component of the total. This is quite consistent with the supposition that the wells sampled are all fed by percolation from the Waimakariri.

Parts per Million Dissolved Solids in various Filtered Waters. — Waimakariri River. Harewood Well. Christchurch Well. Lincoln Well. Total solids 42–42 46–48 80–82 82–86 After ignition 34–36 34–36 54–58 62–66 SiO2 8.0 10.0 14.0 16.0 SO3 2.7 2.7 3.5 3.0 CO211.4 11.4 18.7 22.0 Cl 5.4 4.2 8.5 8.6 N2O50.08 0.3 1.6 0.4 FeO3 0.6 2.0 0.5 0.25 CaO 14.0 12.0 20.0 22.0 MgO 1.5 2.8 4.3 2.9 NaO2 3.0 3.0 1.6 5.7 K2O 2.4 2.4 8.0 6.6 Hardness Grains per gallon— Temporary 1.78 1.73 2.50 3.03 Permanent 0.15 0.26 0.75 0.23 Parts per million— Temporary 25.40 24.80 36.40 43.40 Permanent 2.10 3.70 10.80 3.20 On distributing the acids and bases after the method of Fresenius the following becomes the probable composition of these waters:— Parts per Million in Solution — Waimakariri River. Harewood Well. Christchurch Well. Lincoln Well. SiO2 8.0 10.0 14.0 16.0 Fe2O3 0.6 2.0 * This well was fitted with a pump, through which the water was raised 0.5 0.25 K2SO4 4.4 4.4 7.6 6.5 NaCl 5.6 5.6 3.5 10.7 CaCl2 3.1 1.2 10.4 3.4 CaSO4 1.2 1.2 K2CO3 5.7 4.6 Ca(NO3)2 0.12 0.45 2.4 0.6 CaCO3 21.4 19.0 27.5 36.1 MgCO3 3.1 5.9 9.0 6.1 Total 47.5 49.7 80.6 84.25 Am. nitrogen Not 0.05 0.03 0.25 Organic nitrogen estimated 0.12 0.07 0.05 Dissolved O 7,492 cc. 5,310 cc. 5.460 cc. 6,442 cc.

D. Summary And Conclusions. Observations of the fluctuations of eight wells in and near Christchurch for periods ranging from one to fourteen years show that the wells rise with rain, but the amount of the rise, and the period that intervenes between the rain and the rise, depends greatly on the previous weather. While the wells are raised above normal level by rainfall, they are prevented from falling below normal by percolation from the River Waimakariri; and this is true not only of the town wells, but (contrary to my opinion in 1917) of the Lincoln wells also. The water-analyses are consistent with the thesis that both town and country wells are fed by percolation from the Waimakariri. References. 1. C. Chilton. Some Subterranean Crustacea. Trans. N.Z. Inst., vol. 14, 1881, p. 174. 2. F. W. Hilgendorf. Fluctuations of Artesian Wells near Christchurch. Trans. N.Z. Inst., vol. 44, 1912, p. 142. 3. — Fluctuations of Artesian Wells near Christchurch. Trans. N.Z. Inst., vol. 49, 1917, p. 491. 4. F. W. Hutton. On the Behaviour of Two Artesian Wells at the Canterbury Museum. Trans. N.Z. Inst., vol. 28, 1896, p. 654. 5. R. Speight. Geological Features of Christchurch Artesian Area. Trans. N.Z. Inst., vol. 43, 1911, p. 420. 6. L. P. Symes. Fluctuation of Water-level in a Christchurch Artesian Well. Trans. N.Z. Inst., vol. 49, 1917, p. 493.