This suit was instituted by the United States to enjoin Republic Steel Corporation (Republic), International Harvester Company (Harvester) and Interlake Iron Corporation (Interlake) from depositing industrial solids in the Calumet River without first obtaining a permit from the Chief of Engineers of the Army providing conditions for the removal of the deposits, and to order and direct them to restore the depth of the channel to twenty-one feet by removing portions of existing deposits.

The original complaint, filed November 1, 1954, alleged:

'The defendants, and each of them, are engaged in the manufacture, processing or fabrication of iron, steel and kindred products, and each of the defendants maintains and operates establishments or mills for such manufacture, processing or fabrication along the banks and channel of the Calumet River, Rock Section, between Stations 112 and 200, a navigable water of the United States of America.

'The continued discharge and deposit by defendants, and each of them, of industrial solids and flue dust into the channel of the Calumet River is unlawful and in violation of Sections 403 and 407 of Title 33 of the United States Code. (Secs. 10 and 13 of the Rivers and Harbors Act of 1899, as amended.)

'The continued discharge and deposit posit by the defendants, and each of them of industrial solids and flue dust into the channel of the Calumet River constitutes an obstruction of a navigable water of the United States of America, and constitutes and creates an interference with and an obstruction to interstate and foreign commerce and navigation, to the detriment and injury of the public interest.'

The complaint prayed for a preliminary injunction restraining and enjoining the defendants, and each of them, from depositing and discharging industrial solids or flue dust into the channel of the Calumet River without first obtaining a permit from the Department of the Army providing for suitable and satisfactory conditions for the removal of such deposits and discharges, and for a final decree making the preliminary injunction permanent and commanding and directing the defendants, and each of them, to restore the bed of the Calumet River in front of, or abutting on, their mills or establishments to the original channel depth of twenty-one feet.

On the basis of this complaint, numerous pleadings were filed and the case brought to issue. On Aprial 2, 1956, after the case had been on trial for approximately mately one month, the trial court, over objection by each defendant, permitted plaintiff to file an amended complaint which prayed for the same prohibitory and mandatory relief against each defendant throughout the navigation channel between Stations 105 and 300. Thus, the effect of the amendment was to enlarge the relief sought against each of the defendants to cover the channel for a considerable distance upstream and a longer distance downstream.

On June 19, 1957, the District Court filed its memorandum opinion, together with its findings of fact and conclusions of law. 155 F. Supp. 442. On June 24, 1957, the Court entered a decree providing for both a prohibitory and a mandatory injunction as prayed for in the amended complaint. Defendants were ordered to restore to a depth of twenty-one feet the Calumet channel between Stations 105 and 300, by dredging therefrom industrial solids or wastes including flue dust, the amount of dredging to be done as follows: Republic, 36.78% Of 81.5%; Harvester, 36.31% Of 81.5%, and Interlake, 26.91% Of 81.5,% Total of the dredging done by each of the defendants since 1951.

This decree, on appeal by the defendants, was reversed by this Court with directions to dismiss the amended complaint upon the basis (1) that the acts of the defendants did not violate Sec. 10 or Sec. 13 of the Rivers and Harbors Act of 1899, and (2) that even though a violation had been established, the respective remedies provided by Congress for such a violation were exclusive and did not authorize the relief sought by the government. 264 F.2d 289. Thereupon, the Supreme Court, on petition of the government, allowed certiorari on the issue as to whether the acts of the defendants violated the sections relied upon by the government and whether the remedies provided by the Act were exclusive and barred injunctive relief. The Supreme Court, with four members dissenting, held that this Court had misconstrued the sections in controversy and reversed our decision. 362 U.S. 482, 80 S. Ct. 884, 4 L. Ed. 2d 903. In doing so, the Court stated (362 U.S. at page 493, 80 S.Ct. at page 891):

'Since the Court of Appeals dealt only with these questions of law and not with subsidiary questions raised by the appeal, we remand the case to it for proceedings in conformity with this opinion.'

We think the government in its brief adequately states the contested issues as follows:

'1. Whether there was evidence sufficient to show a continuing violation of Section 13 of the Act of 1899 by discharge of industrial waste into the Calumet River and therefore to support an injunction prohibiting such a discharge.

'2. Whether there was evidence sufficient to show the creation of an obstruction in violation of Section 10 of the Act of 1899 and therefore to warrant an injunction compelling the removal of the obstruction; and, if so,

'3. Whether the allocation of that responsibility between the defendants should be redetermined.'

Thus, a consideration of the record may appropriately be divided into two parts, that which pertains to (1) the prohibitory injunction and (2) the mandatory injunction, which includes the issue as to allocation of responsibility. Defendants state and argue many other issues which, in our view, are subsidiary to the main issues as stated.

Prior to a consideration of the merits, it appears pertinent to make further reference to the reasoning of the Supreme Court. It stated (362 U.S. at page 485, 80 S.Ct. at page 887):

'Our conclusions are that the industrial deposits placed by respondents in the Calumet have, on the findings of theDistrict Court, created an 'obstruction' within the meaning of Sec. 10 of the Act and are discharges not exempt under Sec. 13. We also conclude that the District Court was authorized to grant the relief.'

The Court, after citing and discussing numerous cases, including Sanitary District of Chicago v. United States, 266 U.S. 405, 45S.Ct. 176, 69 L. Ed. 352, still referring to Sec. 10, stated (362 U.S. at page 489, 80 S.Ct. at page 889):

'The teaching of those cases is that the term 'obstruction' as used in Sec. 10 is broad enough to include diminution of the navigable capacity of a waterway by means not included in the second or third clauses. In the Sanitary District case it was caused by lowering the water level. Here it is caused by cloggint the channel with deposits of inorganic solids. Each affected the navigable 'capacity' of the river. The concept of 'obstruction' which was broad enough to include the former seems to us plainly adequate to include the latter.'

Referring to Sec. 13, and particularly the exception of refuse matter 'flowing from streets and sewers and passing therefrom in a liquid state,' upon which defendants relied in the trial court and in this Court as absolving them from responsibility, the Supreme Court stated (362 U.S. at page 489, 80 S.Ct. at page 889):

'As noted, 13 bans the discharge in any navigable water of 'any refuse matter of any kind or description whatever other than that flowing from streets and sewers and passing therefrom in a liquid state.' The materials carried here are 'industrial solids,' as the District Court found. The particles creating the present obstruction were in suspension, not in solution. Articles in suspension, such as organic matter in sewage, may undergo chemical change. Others settle out. All matter in suspension is not saved by the exception clause in 13. Refuse flowing from 'sewers' in a 'liquid state' means to us 'sewage.' Any doubts are resolved by a consistent administrative construction which refused to give immunity to industrial wastes resulting in the deposit of solids in the very river in question.'

Referring to the right of the government to seek injunctive relief, the Court stated (362 U.S. at page 492, 80 S.Ct. at page 890):

'The test was whether the United States had an interest to protect or defend. Section 10 of the present Act defines the interest of theUnited States which the injunction serves. Protection of the water level of the Great Lakes through injunctive relief, Sanitary District (of Chicago) v. United States, supra, is precedent enough for ordering that the navigable capacity of the Calumet River be restored.'

We shall first consider paragraph one of the decree, which relates to the prohibitory injunction. While the findings of the trial court are subject to critical attack, there can be little if any controversy relative to the factual situation upon which the prohibitory injunction rests. That the defendants have deposited in theCalumet River 'refuse matter,' characterized by the trial court and the Supreme Court as 'industrial solids,' is not open to question. Defendants so admitted in their pleadings, but interposed the defense that such deposits consisted of refuse matter within the exception 'flowing from streets and sewers and passing therefrom in a liquid state.' The trial court, sustained by the Supreme Court, held this defense not available. Moreover, at the trial the government as well as the defendants offered extensive proof relative to the extent of such doposits. The dispute here, however, relates in the main to the amount of dredging which each defendant is required to do under that portion of the decree providing for a mandatory injunction. Deposits having been proven as well as admitted, the extent thereof is not material, certainly not decisive as to the propriety of the prohibitory injunction.

The prohibitory decree enjoins and restrains defendants 'from depositing or discharging and from permitting or causing directly or indirectly to be deposited or discharged into the federal channel of theCalumet River, a navigable stream of the United States, * * * between Stations 105 and 300, industrial solids, including flue dust without first obtaining a permit from the Chief of Engineers of the Department of the Army of theUnited States of America, providing for satisfactory conditions for the removal of such future deposite and discharges * * *.' (155 F. Supp. 453). This portion of the decree is based upon Sec. 13 which, so far as here material, provides:

'It shall not be lawful to throw, discharge, or deposit * * * from the * * * manufacturing establishment, or mill of any kind, any refuse matter of any kind or description whatever other than that flowing from streets and sewers and passing therefrom in a liquid state, into any navigable water * * * provided * * * That the Secretary of the Army, whenever in the judgment of the Chief of Engineers anchorage and navigation will not be injured thereby, may permit the deposit of any material above mentioned in navigable waters, within limits to be defined and under conditions to be prescribed by him, provided application is made to him prior to depositing such material * * *.'

Defendants attack this portion of the decree on numerous grounds, most of which have been swept away by theSupreme Court's constructin of the Act. It is argued that the prohibitory decree fails to comply with Rule 65(d) of the Rules of Civil Procedure, 28 U.S.C.A. in that its terms are not specific. It is asserted that the phrase 'industrial solids' is not defined. In view of the history of the case and the opinion of theSupreme Court, there is no merit to this point. Complaint is made that the decree refers to 'a permit' when there is no such document in evidence, not even one in existence. Any merit in this point, however, is clearly dissipated by the plain language of Sec. 13, which provides 'That the Secretary of the Army * * * may permit the deposit of any material * * * within limits to be defined and under conditions to be prescribed by him * * *.' Furthermore, we think a court must assume, contrary to what is argued, that government officials who bear the responsibility will not act arbitrarily or capriciously in the terms and conditions imposed as a prerequisite to the issuance of a permit.

There is, in our view, some valid criticism of the prohibitory decree because of its failure to follow the terms of the statute. Under the statute, the limits and conditions to be imposed as a prerequisite to the obtaining of a permit are lodged with the Secretary of the Army who, in our view, in the issuance of a permit should not be limited by the provision, 'providing for satisfactory conditions for the removal of such future deposits and discharges.' Section 13 authorizes a permit 'whenever in the judgment of the Chief of Engieers anchorage and navigation will not be injured thereby.' It is conceivable that under existing or future conditions it might be determined that some deposits could be made without injury to navigationor the navigable capacity of the river. Under the decree, the Secretary of the Army and the Chief of Engineers would be precluded in such a contingency from the issuance of a permit or from giving consideration thereto.

The prohibitory decree is directed at the making of deposits in the channel of the river. Such deposits are made only through sewers maintained by the defendants at or near their plants, located between Stations 112 and 200. They have no plants, for instance, located between Stations 200 and 300, a distance of almost two miles. Obviously, they have never made any deposits in that part of the channel and never will as long as their plants remain as at present. This may be of minor importance both to the government and to the defendants but, even so, we think that the prohibition against making deposits should be limited to that portion of the river in front of and adjacent to their plants, that is, between Stations 112 and 200.

In our view, and we so hold, the prohibitory portion of the decree should be amended to prohibit the making of deposits between Stations 112 and 200, of industrial solids, including flue dust, without first obtaining a permit from the Secretary of the Army (or his successor in office) within limits to be defined and under conditions to be prescribed by him.

As thus amended, paragraph one, the prohibitory portion of the decree, is affirmed.

Paragraphs 2, 3 and 4 of the decree provide for a mandatory injunction directed at the defendants severally. Paragraph 2 orders Republic 'to restore to a depth of 21 feet below low water datum for Lake Michigan the federal channel of the Calumet River, a navigable stream of the United States of America, * * * between Stations 105 and 300, by dredging within a reasonable time * * * the industrial solids or wastes, including flue dust, deposits by said Republic Steel Corporation in said federal channel of the said Calumet River; the amount of dredging to be done by said defendant, Republic Steel Corporation, shall be determined in the following manner:

'36.78% Of 81.5% Of the aggregate of the industrial solids or wastes, including flue dust, found in the federal channel of the Calumet River aforesaid, and the total of the dredgings of each of the three defendants herein since 1951 in said federal channel, less the total of any dredging the said Republic Steel Corporation may have done in the said federa channel since 1951.'

Paragraphs 3 and 4 are identical with paragraph 2, except taht paragraph 3 is directed at Harvester and enjoins it to dredge '36.31% Of 81.5% * * * less the total of any dredging the said International Harvester Company may have done in the said federal channel since 1951,' while paragraph 4 is directed at Interlake and enjoins it to dredge '26.91% Of 81.5% * * * less the total of any dredging the said Interlake Iron Corporation may have done in the said federal channel since 1951.'

The decree is based upon findings of fact and conclusions of law entered by the trial court following its memorandum opinion. 155 F. Supp. 442. Reference to this citation obviates to some extent the necessity for a lengthy and detailed statement of facts. The record presents a complicated and much confused situation. The government, as the nature of the case perhaps requires, relies in the main upon the testimony of expert witnesses, much of which is unadulterated conjecture, guess and speculation. Many of the trial court's findings, some of which are crucial, are based upon such testimony. Any attempt on our part to enumerate thoe which are supported by legitimate opinion testimony and those predicated solely on speculation and conjecture would constitute an intolerable burden and perhaps in the end serve no good purpose. In any event, the conclusion which we have reached makes such an attempt unnecessary.

Even so, some statement of facts leading up to what we regard as the decisive issues appears to be in order. The Calumet River is a busy waterway and is used by a large number of lake and foreign ships, many of which are 600 or more feet in length, with a 60-foot beam and a loaded draft of 21 feet and propeller diameter of 15 feet. The river extends from Lake Michingan mostly in a southerly direction and originally emptied into the lake. This flow, however, was reversed in 1922 in order to divert the polluted waters away from the lake. Thus, with rare exceptions, the normal direction of flow of the river since 1922 has been away from the lake. The level of the lake varies the flow in the river and, when rising, increases its velocity.

The Corps of Engineers of the Department of the Army has been designated by Acts of Congress to maintain a federal channel in the river for the purpose of navigation, and has done so since 1881. The river channel is marked at 100-foot intervals called 'Stations.' Station 0 is near the Elgin, Joliet and Eastern lRailroad bridge and Station 300 is just north of 130th Street, a distance of more than five and one-half miles. Station 0 also marks the point where the water from the lake enters the river channel which from Station 0 to Station 300 has a minimum width of 200 feet, with a disignated depth of 21 feet.

All three defendants have been engaged for many years in the production of iron and relatd products, in which they use iron ore, limestone and coke. Harvester is located on the west bank of the river, with a river frontage extending from about 108th to 111th Street. Interlake is located on the east bank of the river from about 107th to 111th Street, and also on the west bank of the river from 111th to 114th Street. Republic is located on the east bank, with a river frontage extending from 111th to about 112nd Street.

Republic produces approximately 500,000 gross tons of iron and 1,300,00 gross tons of steel ingots per year; Harvester, approximately 650,000 gross tons of iron and 1,000,000 gross tons of steel ingots per year, and Interlake, approximately 450,000 gross tons of iron per year but no steel ingots. As a result of this production, certain industrial wastes are created, consisting of flue dust containing a high percentage of iron oxide, slag, calcined lime and other solids. Other wastes such as coke, breeze, mill scale and other solids are created from numerous plant operations.

Defendants pump water in vast amounts out of the river for use in their manufacturing processes for cooling, removing solids from industrial gases and transporting industrial wastes. After use, the water is returned through the various sewers of the defendants which empty directly into the river or into two slips which open into the river. Republic has eight sewers, Interlake five and Harvester fourteen, a total of twenty-seven sewers, emptying into the river directly or by way of said slips.

The water when pubped from the river is not fit for some uses because of its dirt content. For instance, Republic uses some 50,000,000 gallons of water per month, which it obtains from the City of Chicago. Some 15,000,000 gallons of this water are used for sanitary and the remainder for manufacturing purposes. Republic takes its water from the river by means of two pumps and its principal use is for cooling purposes. There is evidence that the water pumped each month by Republic from the river contains 239 tons of solids. There is similar evidence regarding solids in the water pumped from the river by the other two defendants. The Court made no finding as to the quantity of solids contained in the water when pumped from the river. It did find, 'These solids (referring to those deposited) are in excess of those being transported in the water at the time it is being pumped out of the river by the defendants.' (155 F. Supp. 450) The Court, in determining the responsibility of the respective defendants, neither in its findings nor in its decree made any allowance for industrial wastes in the water at the time it was pumped from the river. (Further reference will be made to this point.)

The Court in its findings recognized the effort made and means employed by the defendants to remove industrial solids and flue dust from water before its return to the river but did not indicate the measure of their success in this respect. While their methods vary to some extent, they are designed to and do accomplish substantially the same results. Harvester claims that its method is more than 98% Efficient in removing industrial solids from the water before it is returned to the river; Republic contends that its method is 97% Efficient, and Interlake contends that its method is more than 99% Efficient. The soundness of these contentions is not disputed by the goverernment; in fact, it states in its brief, 'Despite the 97% And 98% Efficiency of these thickeners some solids get through them.'

In addition to the twenty-seven sewers of the three defendants, there are thirty-seven other sewers which discharge into the Calumet River between Stations 0 and 300. The government did not take or analyze a single sample from any of these other sewers. Defendants introduced analyses of samples taken at the sewers of six other industrial plants on the river, each of which contained solids, some in substantial amounts. Among the other thirty-seven sewer outfalls are three of United States Steel Corporation which are 40', 30' and 18' in size, and two of Youngstown Sheet and Tube Company which are 120' and 136' in size. Both of these plants are located near Station 0, some two miles upstream from the plants of the defendants. The balst furnace capacity of United States Steel is several times larger than that of the three defendants combined. The blast furnace capacity of Youngstown is larger than that of any of the defendants. A government witness admitted that material coming from the sewers of these two industries might move down the river to the area in front of defendants' plants and as far south as Turning Basin No. 5, near Station 300. Also located upstream from defendants' plants, near Station 50, is Commonwealth Edison Company which, according to estimates, deposits in the Calumet River about 2,190 tons of solids per year. Other industries located on the river, downstream from defendants' plants but within the area defendants are required by the decree to dredge, are Great Lakes Carbon Corporation, Globe Asphalt Company, General Chemical Company, Allied Chemical Company and Ford Motor Company. All of these industries deposit some industrial waste in the river.

The propriety of a mandatory as well as a prohibitory injunction (the latter heretofore discussed) is simplified by the fact that it is without dispute, in fact admitted by defendants, that each has been and continues to make deposite of industrial solids in the river channel, which theSupreme Court has held is unlawful and in violation of the Rivers and Harbors Act. Further, defendants do not dispute, in fact they concede, that they are severally responsible for the removal of such deposits, less the amounts which they have by dredging removed since 1951. (Interlake contends that during that time it has removed more than it has deposited and therefore has no responsibility to do further dredging.) Defendants assert and the government does not dispute but that the three defendants in 1951 dredged the channel in a manner satisfactory to the Army Engineers.

There runs through the trial court's reasoning and the government's brief the thread of though that because defendants have violated the Rivers and Harbors Act they are subject to a penalty. For instance, the trial court in its memorandum opinion, referring particularly to Harvester, stated that it, 'like the other two defendants, has required the Government to come into court and point out just exactly what it should remove from a public stream though it knows, or ought to know, better than anyone else just exactly what it put in the stream. Since the Government was required to come into court and spend long weeks in litigation, it seems to the court that it is entitled to complete relief, including a mandatory injunction * * *.' The government in its brief states:

'As the Supreme Court has now made clear, this is a case of enforcement of the federal statute and, so far as the mandatory injunction is concerned, is a case of compelling correction of conditions which resulted from violations of the statute.'

Thus, it appears that the trial court at the government's insistence has penalized the defendant for resisting the government's action for equitable relief. There are numerous fallacies in such reasoning. At the moment, we note (1) that defendants did know and prove the maximum quantity of industrial wastes which they severally had deposited in the river, which proof was rejected by the Court, and (2) that the government was not relieved of the burden of proving the case which it had alleged, that is, the several responsibilities of the defendants. More than that, we think there is no basis for the implication that the defense was not made and carried on in good faith. It is true that the defendants in 1951 dredged all material, industrial and otherwise, from the channel. This was done, however, without prejudice to their rights in any future controversy with the government. The record supports an assertion by defendants that this long and expensive litigation is the result of the government's insistence tht they assume the responsibility of dredging 100% Of all material in the channel, industrial or otherwise. Even the decree under attack does not impose such a responsibility. It is also pertinent to note that defendants defended their activities on the ground that they were not in violation of theRivers and Harbors Act and, in any event, the Court was without power to award injunctive relief. There is no basis for the imputation that this defense was not logical or interposed in good faith, in view of the fact that it was sustained by this Court as well as by the four dissenting members of the Supreme Court.

As a result of our study of the record and the briefs of the respective parties, we find it unnecessary to discuss all the questions which have been argued. This is for the reason that the undisputed evidence, including that offered by the defendants in connection with the decision of the Supreme Court, leaves only the several responsibilities of the defendants as the fundamental and controlling issue. We conclude that such responsibilities have been erroneously determined and that the portion of the decree relative thereto must be reversed. We shall attempt to justify that conclusion.

In the beginning, it must be kept in mind that this case was pleaded and tried on the theory that defendants were severally responsible for idustrial solids and flue dust deposited in the channel of the Calumet River. To determine that responsibility, the government with its expert witnesses relied upon analyses of samples taken with the Petterson 'Grab Sampler' from the bottom of the river, certain slips and two turning basins. The bottom samples were taken from fourteen different locations scattered between different locations scattered between Station 20 and Turning Basin No. 5, a distance of over five miles. Eight of the twenty-two samples were taken in February 1955, and fourteen inNovember 1955. Eight of the November and the eight February samples were taken at the same locations. Some samples were taken in slips and in Turning Basin No. 3. The government also took some samples from defendants' sewers. Both the sewer and the river bottom samples were analyzed chemically and petrographically for the purpose of determining their components. As we understand, the government relied upon an analysis of sewer samples for the purpose of showing their nature but not the quantity of material contained therein. It relied upon the analyses of bottom samples for the purpose of determining the quantity and location of industrial solids in the channel and that a portion thereof was of the same general nature as that found in the flow from defendants' sewers. It is not claimed, however, in fact the record shows to the contrary, that such analses proved the source of the deposits.

Defendants sought to prove their responsibility by measuring suspended solids in sewer effluents. Measurements were conducted duringFebruary and March, 1955, by J. William Fehnel, a consulting engineer, familiar with the measurement of industrial wastes. Samples of defendants' sewer outfalls were taken at different hours of the day over a period of several weeks. Samples were also taken by each of the defendants at their intake pumps during a similar period. In all, Republic took some 432 samples; Interlake, 240 samples, and Harvester, 308 samples. The samples were composited and analyzed for total solids. From the analyses of these samples defendants' expert witnesses calculated the quantity of industrial wastes which flowed from their sewer outlets into the river, and also the quantity of industrial solids taken into their plants with water pumped from the river. Defendants' calculation as to the amount of industrial wastes deposited by the respective defendants was the maximum amount for which, according to their theory, they could possibly have been responsible. In this connection, it is pertinent to observe that some and perhaps much of the deposits remained suspended and were carried away with the stream. As we understand, only the deposits which settled did or could have contributed to any obstruction or affected the navigable capacity of the river.

The two principal witnesses relied upon by the government in support of the decree allocating responsibility were Chester R. Pelto,Chief of thePetrographic Section of the Divisional Laboratory for the Missouri River Division of the Corps of Engineers, and Hans Albert Einstein, a Professor of Hydraulic Engineering at theUniversity of California. Actually such allocation depends mostly upon the testimony of Dr. Einstein. All objections to the testimony of these two expert witnesses were overruled and they were given a free rein to guess, conjecture and speculate.

Pelto testified that his section made petrographic analyses of samples sent to it from Chicago in 1955. These bottom samples in his opinion evealed a high content of material not natural to the Chicago area. Exhibit 171, consisting of two graphs drawn by the witness to illustrate the result of his analysis of steel mill effuent constituents, was introduced and explained. Exhibit 172, also prepared by the witness, entitled 'Chemical Analyses of minus 325 mesh solids from Calumet River channel samples,' was introduced and explained. This exhibit depicted two curves showing the amount of material which the witness estimated to be of steel mill origin.

We have read and reread the testimony of Dr. Einstein and are much impressed with his acumen in both the engineering and legal fields. He demonstrated a keen understanding of the issues as well as the result which the government sought to achieve and quite plainly accommodated his testimony accordingly. He analyzed no samples, in fact had little if any personal knowledge of the river, the channel or their contents. His testimony was based upon exhibits prepared by other witnesses and introduced by the government.

Dr. Einstein to illustrate his testimony prepared government Exhibit No. 182, a grahp purportedly showing the percentage of the material in the river bottom which was finer than 44 microns. The graph was based on the table in a government exhibit previously prepared by Pelto, which in turn was based on the bottom samples. Einstein also prepared and identified government's Exhibit No. 183, which purports to show the iron oxide accumulation for a ten-year period, from 1946-1955, in cubic yards per lineal foot. This exhibit was also based on a government exhibit previously offered in evidence. He thought from the deposits depicted in Exhibit 183 that much of the iron oxide must have flocculated. He described at length his theory of flocculation as a process by which tiny particles of material come together and form larger particles which have a tendency to settle. He concluded that the large amount of iron disposition was in the area near defendants' plants. He stated, 'The only possibility that exists to make that material settle out within a very short distance from its point of origin is for it to be flocculated.' He concluded his dissertation on his flocculation theory with a statement which bears quoting:

'So you see that responsibility in the case of material like that is not something that is uniquely defined by just the physical aspects of the final product; and for that reason,I believe one of the most important ideas and aspects that will enter this case is to actually define how to determine responsibility.

'I think this is the crux of the whole matter. It means we must define how we want to assess responsibility, and to my thinking, the biggest difference between the results that the different parties in this case have obtained is caused by the difficulty by which they are assessing the responsibility.

'I believe the basic idea is that this case ought to teach us how to assess the responsibilities.'

While we think the record clearly reveals that theplace where deposits would settle in the channel with reference to the place of their deposit would depend upon the velocity of the stream, yet Dr. Einstein testified:

'Q. Now, in you calculations, what velocities did you have in mind, Professor? * * * A. I did not actually use, in that calculation, the value of the velocities at all.

'Q. Would a change in velocity make any difference? A. Very much so. * * * The only reasonable way of finding out, that I know of, is to make an experiment, that means, to actually try what velocities are necessary to remove this muck.

'Q. And you have not done that?

A. no, I have not.'

With reference to the particles which he though would flocculate, he testified:

'Q. Have you ever examined any of these particles under a microscope? A. I have seen particles just like it under a microscope but I have not examined these particles. However, there are lots of particles in there that are too small to be seen under a microscope and those particles are the worst offenders because they make the strongest and the fluffiest of the deposits. * * *

'Q. Do you know how fine flue dust is, Professor? A. No.

'Q. Do you know whether or not it will flocculate? A. I have not made any experiments with what you call flue dust.'

Thus, under Dr. Einstein's theory, most of the shoaling was caused by the flocculation of tiny particles less than 44 microns in size and 'the worst offenders' were those which could neither be filtered nor seen with an ordinary microscope. It seems strange that these 'worst offenders' which are so much relied upon by Einstein in his flocculation theory were never seen by him and, so far as this record discloses, had never been seen or heard of by any other person. Harvester in its brief, referring to this point, states:

'It is submitted that if Professor Einstein and the Army Engineers had any real confidence in the theory that invisible particles were 'the worst offenders', they could have made some examinations with an electron microscope, which would have shown such particles if they existed.'

We think that is a pertinent observation, particularly in view of the fact that Dr. Einstein late in the trial, referring to 'material of less than one micron,' admitted, 'If you use X-ray or if you use an electron microscope, then you can see it very easily, and that is the reason people use these means, because finer particles won't show with these waves.' The government makes no explanation for its failure to prove that the so-called flocculated material was inexistence.

Another factor which, in our view, seriously detracts from Einstein's calculation as to defendants' responsibility is that no consideration was given to the water content at the time it was pumped into their plants from the river. Pelto, upon whose charts and grahs Einstein relied, assumed that the water pumped into defendants' plants carried no suspended solids and that all material found in the effluents was chargeable to defendants even though it could not be identified as waste from their mills. We need not relate the testimony orrered by defendants as to the large amount of industrial solids contained in the water when pumped from the river because the fact is not in dispute. As previously noted, the Court found:

'These solids are in excess of those being transported in the water at the time it is being pumped out of the river by the defendants.' (155 F. Supp. 450).

The Court interrogated Dr. Einstein on this point, calling his attention to Republic's proof showing that 34,745 tons of industrial wastes were discharged from its sewers during the years 1951 to 1955, and that during the same time 11,070 of those tons were in the water pumped into its plant. The Court inquired, 'Do you think it is right to substract that 11,070 from the 34,745 to determine the amount of solids that might be deposited in the river?' The witness answered, 'As long as solids go away from the river, they must be subtracted. I think the procedue is proper all right.' Of course, Dr. Einstein was referring to the effluent test method proposed by the defendants, but if they were entitled, as we think they were, to deduct the amount of industrial waste in the water which they pumped from the river from that which they discharged, under their theory of determining responsibility, it is not discernible why they would not be entitled to the same consideration under the theory employed by the government and adopted by the Court. The government in its brief appears to argue otherwise. It states, 'Thus, the fact that some solids are in the water when defendants take it out of the Calumet does not mean that they are free to put them back.' That may or may not be true but, assuming it is, it does not follow that in assessing their responsibilities for the creation of an obstruction they were not entitled to credit for the amount of industrial wastes which they pumped from the river.

Einstein's theory on flocculation presents an intriguing and interesting subject for speculation. He admitted it was based on the assumption, without proof, that flocculation actually took place. Even so, he guessed that the particles were there, and the government, as well as the trial court, attached much weight to the verity of the theory. The government in its brief states:

'In ignoring these fine particles, defendants wee ignoring the effect of flocculation, a description of particle behavior which the trial court accepted. Even if it were possible to determine exactly the amount of solids being discharged from any one plant, the occurrence of flocculation would be an additionl factor in assessing responsibility for obstruction. * * * So it would be not only the materials discharged by a party that would determine how much of an obstructon he has created, it would also be the materials whose progress through the river has been arrested by his flocculated deposits.'

Assuming that this flocculation theory has any evidentiary support, which it does not, the question is raised as to whether it is outside the issues made by the pleadings. Defendants were charged with making deposits of industrial solids and the government asserted responsibilities based on that premise. The case was tried on the theory that defendants were responsible only for deposits actually made; in fact, the Court found that industrial wastes, including flue dust, were deposited in the river by the three defendants. (As shown, Einstein testified that he did not know if flue dust would flocculate.) We need not dwell further on this phase of the case because whether the government's theory be accepted or rejected is relatively unimportant in considering defendants' responsibilities as decreed by the Court.

The Court rejected defendants' proof of analysis of their sewer effluents as a means of determining the quantity of their deposits. The government in its brief states:

'But the meaningfulness of sewer sampling over a short period of time was one of the hotly disputed issues of the trial, and one government witness expressed the opinion that sewer sampling for anything less than a period of years would not give an adequate indication of the nature of the effluent. The trial court, as was its prerogative, apparently chose to give this opinion the greater weight, as is indicated by the statements in Finding 31 that any sewer sample is 'actually only representative of what is in the sewer effluent at the time the particular sample is taken."

This statement is subject to criticism in numerous respects. True, one government witness expressed the opinion as stated; however, another government witness, as well as a number of defendants' witnesses, expressed the opinion that sewer sampling was the recognized method of determining the quantity of solids carried into the river. The fact is that the government relied upon a few sewer samples for the purpose of showing the nature but not the quantity of industrial wastes which defendants and deposited in the river. It is also pertinent to note that the government relied upon its bottom samples taken at fourteen different locations over a five mile stretch of the river as being sufficiently representative to show both the nature and quantity of industrial solids in the river channel. We are not able to reconcile that reliance with the rejection of defendants' contention that several hundered sewer samples were not sufficiently representative to show the nature and quantity of deposits which they had made. In view of the fact that the case is to be reversed for a new trial, we need not dwell further on this point. However, keeping in mind that the primary issue is as to the nature and quantity of deposits which defendants have made in the river, a commonsense view strongly suggests that the ascertainment be made at the point where such deposits take place. That approach goes directly to the problem, in contrast to the government's conjectural and speculative theories based upon river bottom samples.

'Of the industrial wastes, including flue dust, in the 21 foot Federal channel of the Calumet River between Stations 105 and 300, not less than 81.5% Thereof was deposited therein by the three defendants.'

The basis for this finding is revealed in the Court's memorandum opinion, in which it is stated:

'There is evidence that of the shoaling of the river between Stations 105 and 300, resulting from the deposits of industrial waste therein, the three defendants together are responsible for approximately 70% To approximately 93% Thereof.'

These calculations are based upon the testimony of Einstein but even he did not testify that defendants were responsible 'for approximately 70% To approximately 93%.' He expounded different theories for determining responsibility. On one theory he fixed it at 93% And on another at 70%. The Court in its memorandum stated:

'The court, upon consideration of all the evidence, has reached the conclusion that the mean of the aforesaid estimates, that is, 81.5% Fairly represents the aggregate responsibility of the three defendants * * *.'

Thus, the Court, in fixing the aggregate responsibility, split the difference between Einstein's two estimates.

Einstein's first method was to determine the amount of suspended matter discharged from defendants' sewers. He rejected this direct approach for the principal reason that, in his opinion, it did not take into account the different volumes which the finer and coarser particles, respectively, would assume upon reaching the river bottom. The second method was to 'take all the materials which have been contributed to the river bottom by other sources,' and subtract that amount from the total existing deposit volume so as to determine the reposit of the defendants. His third method was 'more or less an inbetween method,' and was to take the total existing deposit volume and divide it up, 'according to some kind of key into two parts,' one part being the responsibility of the defendants.

The government's effort to support the twofold calculations by Einstein as to defendants' aggregate responsibility is anemic. In its brief it states, 'Both the 70% And 93% Figures appear in Dr. Einstein's testimony.' The fact that they appear does not support the finding as to aggregate responsibility. In arriving at the 70% Estimate, Einstein relied upon figures submitted by Pelto, which he admitted were quite erratic. His 93% Estimate was made by estimating all material in the channel of the river, by estimating contributions from other sources and assigning the remainder to defendants. In other words, he estimated the deposits from all other sources at 7%, leaving defendants responsible for the remaining 93%. Einstein's speculation that of the deposits in a five mile stretch of the river only 7% Was attributable to other sources borders upon the preposterous. As we have previously shown, there were many other industries located on the river, with water being discharged from thirty-severn sewers other than those of the defendants. Particularly, there were United States Steel Corporation and Youngstown Sheet and Tube Company, both located upstream from the plants of defendants. They were engaged in operations similar to those of the defendants, with a potential for the discharge of waste matter in the river far greater than that of the defendants. The witness admitted on crossexamination that in making his 7% Estimate he did not take into consideration 550,000 cubic yards put into the entrance of the river in the past ten years by United States Steel Corporation. He was asked how he estimated the number of yards from the sewer outflow of Commonwealth Edison Company and he answered, 'I would definitely say that what I put down there was a guess that I made on the basis of the information that I got from the testimony.'

It is true the Court in its findings, following the government's theory, minimized the quantity of deposits from sources other than defendants. The Court found: $510'Since 1951, these two companies (United States Steel and Youngstown) have paid for the dredging annually of all shoals in the entrance channel.' (155 F. Supp. 450).

There is evidence that these companies dredged in front of and adjacent to their plants in the spring of each year but this does not explain what became of their deposits between the times of dredging. The Court also found:

'Those shoals were composed of steel mill wastes but the evidence shows that the solids composing those shoals did not travel downstream as far as the plants of the three defendants.'

It is a phenomenon, we think unexplained, as to why industrial wastes deposited in the river by United States Steel and Youngstown would not travel downstream as far as defendants' plants, a distance of less than two miles, while at the same time industrial wastes from defendants' plants would travel downstream a distance of more than three miles. More than that, government witnesses, incluiding Einstein, admitted taht at Station 300 there were probably industrial wastes in the river which had been discharged by United States Steel and Youngstown.

'The presence in the record of detailed testimony concluding that defendants contributed 70% Or 93% Of the sedimentation in the river is sufficient to support a finding that defendants are responsible for either of those percentages or any percentage in between.' Thus, so it is argued, 'The District Court's conclusion that the three defendants were responsible for at least 81.5% Of the deposits is supported by evidence.' In our view, this is a dubious argument, particularly in view of the wide variance in the percentages estimated by Einstein. It seems to us that each of these estimates seriously impairs any weight which could properly be attached to the other. If a witness at one session of court testifies that the age of a horse is ten years, and at another session that its age is six years, we do not believe that such testimony would support a finding as to either, and certainly not a finding that its age is eight years.

'The Court's action in this case is analogous to the action taken by triers of fact in a wide variety of valuation cases, such as condemnation cases.'

Cited as a typical example is United States v. 2.4 Acres of Land, More or Less, 7 Cir., 138 F.2d 295, 297. This case does nothing more than recognize the usual rule in condemnation cases that where the finding as to value is within the scope of the expert testimony it will not be distrubed. Neither this case nor any other, so far as we are aware, presents a situation where the same 'expert' testified as to two separate values, the higher being more than 30% Above the lower.

Dr. Einstein, near the conclusion of his cross-examination, was asked, 'But how do you know the solids in the channel came from the sewer effluents?' He answered, 'Well, that is the $64,000 question.' The answer thus given to this vital questiona is as appropriate today as it was at that time. In our judgment, the finding that of the industrial wastes, including flue dust, in the channel of the Calumet River between Stations 105 and 300, 'not less than 81.5% Thereof was deposited therein by the three defendants,' is without evidentiary support and clearly erroneous.

While we are satisfied with the soundness of this conclusion, we recognize the possibility that we might be in error. For the purpose of the discussion which follows we assume, contrary to what we have held, that the aggregate responsibility of the three defendants was as determined.

This brings us to the finding relative to the several liabilities of the defendants. The Court found that of the aggregate deposits made by the three defendants, Republic deposited 36,78% Thereof, Harvester, 36.31%, and Interlake, 26.91%. Pertinent to this finding is the question as to whether the liability of the defendants was several, or joint, or joint and several. At the conclusion of the government's case, all defnedants moved to dismiss on the ground that the complaint alleged that the liability of the defendants was several and that the government had offered no proof in that regard. The Court reserved ruling on these motions and in its memorandum opinion stated:

'The Government has not in this case contended for joint liability and, accordingly, the court will not decree joint liability. Had the Government contended for joint liability, and had the case been tried on that basis, the court is inclined to think that it would have held the defendants jointly and severally liable for the commission of a public nuisance by obstructing a navigable stream.'

The government in its brief concedes that it did not allege joint liability because this is not a tort case but one for the enforcement of a federal statute. The government states:

'Certainly a statutory violator should not be permitted to escape the effects of his violations because they are mingled with the effects of other violations in the same area. Thus, if the phusical conditions require removal of deposits made by others to cure the statutory violations, that is a responsibility defendants assumed when they violated the statute. Hence, if the 81.5% Or other allocation of responsibility should be held to be unprovable, the injunction should order defendants to remove all deposits rather than to escape the effect of their violations.'

This argument, if sound, which we think it is not, means that the government is entitled to the relief afforded by the mandatory injunction, irrespective of whether it has proven several liability. In fact, it is an argument in favor of joint and several liability, contrary to the action alleged in the complaint as well as the theory on which the case was tried and decided. The argument, if accepted, means that the government could have proceeded against any one of the defendants or against any other manufacturer making deposits in the river, however small, and obtained a decree requiring such defendant to dredge all material from the river because its deposits had become intermingled with those made by others. The argument on its face is fallacious.

The record reveals that the trial court at the conclusion of the case was much disturbed as to whether there was any proof of individual responsibility. The Court stated:

'If there is in the record sufficient data and information available to determine the defendants' contribution of industrial waste into the Federal channel, I would like to have that data pointed out to me. * * * you must point out to me the data and the evidence which indicates the respective contribution of the defendants to the waste which is said to be in the channel. If you can't do that, then you had better put some evidence in.'

The Court, on December 12, 1956, entered a formal order requiring the government 'to point out to the Court the data and evidence which indicates the respective contributions of the defendants to the waste which is said to be in the channel * * *.'

On June 24, 1957, more than six months later, the government without offering any further testimony and for the first time came up with a theory, apparently embraced by the trial court, in support of the finding fixing the percentage liability of each of the defendants. The theory was based not upon government proof but upon that of the defendants relative to their effluent tests offered for the purpose of showing the maximum amount of deposits which they had severally made. That testimony, as previously stated, was rejected by the Court as a proper means of showing the volume of deposits for which defendants were responsible.

The government is now in the anomalous position of defending such rejection but at the same time relying upon defendants' evidence for the purpose of showing their several responsibilities. As the findings disclose, such testimony was rejected because the sewer samples relied upon by the defendants 'vary greatly from time to time' and are 'only representative of what is in the sewer effluent at the time the particular sample is taken.' If that is correct, as the government asserts and as the Court found, we are not able to discern how it can support the finding which allocates a specific percentage of responsibility to each defendant. We think such finding is clearly without any substantial support.

The government concludes its argument in support of the mandatory decree:

'* * * we should note that the answer to the claim that the injunction is indefinite because 81.5% Of the deposits cannot be identified is that the river will have to be dredged and when that is done 100% Will be identified and determination of 81.5% Then becomes a matter of simple arithmetic.'

In our view, this is a remarkable and highly confusing statement. It appears to mean that the government has abandoned not only the theory upon which the case was tried but the mandatory portion of the decree which determines both the aggregate and several responsibilities of the defendants. Boefore aggregate responsibility can be determined the river will have to be dredged of 100% Of all material, industrial solids, flue dust and otherwise, irrestpective of by whom deposited. Does the government mean that this 100% Dredging will have to be done by the defendant? If so, is the responsibility joint or several or both? Who is to compensate for the dredging of material for which defendants admittedly have no responsibility? It must be a novel theory that a party charged with an obligation under a mandatory decree must perform more than required in order to ascertain the extent of its obligation.

We have considered numerous other questions which have been argued but we need go no further. We hold that the government was entitled to the prohibitory injunction, amended as previously suggested, and that it was not entitled to the mandatory injunction as decreed.

'All that could be tried in the new trial that all defendants ask for is the matter of allocation of responsibility. In other words, a new trial could only determine a new way of splitting up the costs of dredging the river.'

The matter of allocation of responsibility, however, is the crux of this case and a new trial must be had for that purpose. It cannot be avoided on the basis that it 'could only determine a new way of splitting up the costs of dredging the river.'

We are keenly aware of the diffcult problem with which the government is faced, as well as the burden which a new trial will impose upon a trial court. Even so, we cannot avoid the conclusion taht the government the same as any other litigant in this or any other case cannot escape the burden of proving the case which it has alleged and of sustaining the theory on which the case was tried and decided. This the government has failed to do.

In accordance with what we have held, the decree is in part affirmed and in part reversed, and the cause is remanded for a new trial.