The petitioners, representatives of the corn wet milling industry, seek direct review of regulations promulgated by the Environmental Protection Agency setting forth standards of effluent discharges for new plants in this industry under § 306 of the Federal Water Pollution Control Act Amendments of 1972. 33 U.S.C. § 1316.

The new plant standards are before us for the second time. In CPC International Inc. v. Train, 515 F.2d 1032, 1034 (8th Cir. 1975) (CPC I ), we held that this Court had jurisdiction to directly review the standards for new plants under § 509(b) of the Act. We also held that this Court did not have such jurisdiction with respect to existing plant guidelines, a position to which we continue to adhere. Id. at 1036-1043.1 We also held that there was not sufficient evidence in that record to support the new plant standards promulgated by the EPA and remanded the matter to it with instructions to either:

furnish support for the new source standards previously published, or establish new ones which can be achieved with the best available demonstrated control technology.

Id. at 1050 (footnote omitted).

We also instructed the EPA to update its projected cost figures.

On remand, the EPA undertook a review and reconsideration of the new plant standards. Following appropriate administrative procedures, the EPA resubmitted the standards previously promulgated.

(1) the EPA failed to give a fair and reasoned consideration of its new source standards and instead prematurely and arbitrarily adopted its original standards;

(2) the EPA erred in determining that the 1977 guidelines can be met by new plants utilizing proposed 1977 technology;

(3) the EPA erred in holding that the addition of deep bed filtration will enable industries to meet the proposed new source standards; and

(4) the EPA erred in determining that the suggested technology could be acquired within the cost figures projected by it and in determining that the industry could afford to build new plants incorporating the technology necessary to achieve the new source standards.

I. The EPA Acted Properly on Remand.

We are not entirely satisfied with the record made by the EPA on remand. Much of the data in the record is presented without careful collation, evaluation and simplification. Much of the statistical information is quantified in units of measurement other than those used in the proposed standards, pounds per thousands of bushels, thus making it difficult for us to determine whether the data supports the EPA's conclusion. Expert opinion is frequently unsupported and little, if any, effort is made to make scientific testimony understandable.

Nonetheless, we cannot say that the new source standards should be set aside in their entirety on the grounds that the EPA prematurely and arbitrarily rubber stamped the previously approved standard. It did substantially more than this. It retained expert consultants to assist in the decision-making process, it undertook a review of the experience of other municipal and industrial users in the United States and abroad, it undertook a review of the literature in the field, and it made a serious and a partially successful effort to obtain all available information from presently operated corn wet milling plants.II. The Record Supports the Conclusion of the EPA that the 1977 Existing Plant Guidelines can be met by New Plants Utilizing the Proposed 1977 Technology.

The petitioners, in effect, seek to relitigate issues decided in our first opinion. In CPC I, we held that:

(1) the new source standards are identical to the 1983 guidelines for existing plants;

(2) the new source standards/1983 guidelines "are predicated on the availability of the 1977 technology plus the addition of deep bed filtration;"

(3) the EPA did not err "in determining that the 1977 technology, when employed in a new plant, would enable it to comply with the 1977 guidelines;" and

(4) the EPA adequately considered contentions of the industry pertaining to "shockloads," "variability" and "excursions" in determining the 1977 existing plant standards. CPC International Inc. v. Train, supra at 1045, 1046 and 1046 n.30 (footnotes omitted).

The petitioners suggest that the data presented initially and on remand does not support the EPA's conclusion that the 1977 existing plant guidelines, upon which the final new plant standards are based, can be met. They argue that the EPA has failed to adequately consider the raw wasteload of the corn wet milling industry, the variation of the wasteloads according to the particular product produced, the temporary "shockloads" of waste created upon shifts from one product process to another or the "excursions" or high waste level concentration that occurs in this industry from time to time.2

We have carefully reviewed the entire record and remain convinced that new plants constructed with proper in-plant controls and an adequate activated sludge or similar treatment facility can meet the 1977 existing plant guidelines of removing all but 50 pounds each of BOD 5 and TSS per MSBu.3 The record discloses that two existing plants without all of the available technology proposed for use in a treatment facility or appropriate in-plant controls have met or nearly met the 1977 guidelines. At the American Maize plant, the monthly figures of January through March, 1975, are below the 1977 standards. At CPC-Corpus Christi during the period of December, 1974 through April, 1975, the monthly BOD 5 effluent level is just above the maximum thirty-day period 1977 guidelines. These plants have achieved this level of effluent reduction even though they have been inhibited in their efforts by space and design limitations. New plants are being designed to avoid such limitations.

We recognize that the District Court for the Southern District of Iowa in Grain Processing Corporation v. Train, 407 F. Supp. 96 (S.D.Iowa 1976), held that the 1977 guidelines cannot be met by existing plants. That decision is not before us here. We are concerned only with the standards for new plants. The District Court's decision is on appeal to this Court and will be considered in due course. We note only that a new plant can incorporate in-plant controls and surface condensers with moderate expense and inconvenience while existing plants have more difficulty in converting to this technology.

We affirm our decision in CPC I that the EPA did not act arbitrarily or capriciously in concluding that the 1977 technology employed in a new plant would enable it to meet the 1977 existing plant guidelines.

III. The Proposed New Plant Standards for BOD 5 can be met; the Proposed New Plant Standards for TSS cannot be met.

Section 306(a)(1) of the Act, 33 U.S.C. § 1316(a)(1), states that the new source standards are to reflect:

(T)he greatest degree of effluent reduction which the Administrator determines to be achievable through application of the best available demonstrated control technology, processes, operating methods, or other alternatives, including, where practicable, a standard permitting no discharge of pollutants.

Our task is to determine whether the EPA's decision is "arbitrary, capricious, an abuse of discretion, or otherwise not in accordance with law." 5 U.S.C. § 706(2)(A); Citizens to Preserve Overton Park v. Volpe, 401 U.S. 402, 416, 91 S. Ct. 814, 28 L. Ed. 2d 136 (1971); CPC International Inc. v. Train, supra at 1044.

The EPA has proposed that a new plant in the corn wet milling industry shall achieve an average of daily values for thirty consecutive days that shall not exceed an effluent level of 20 pounds of BOD 5 and 10 pounds of TSS per MSBu. This standard is to be reached through the use of a complete activated sludge system to reduce the wasteload to 50 pounds each of BOD 5 and TSS per MSBu, the 1977 guidelines, and a subsequent reduction of 30 pounds of BOD 5 and 40 pounds of TSS per MSBu through the addition of a deep bed filtration system. The EPA asserts that the experience of the Clinton corn plant treatment facility, which includes the use of deep bed filtration, supports the new source standards. It contends that deep bed filtration is a demonstrated technology in the treatment of wastewater in municipal and industrial plants and its use has resulted in effluent discharges containing essentially no suspended solids. It finally asserts that deep bed filtration technology can be transferred to the corn wet milling industry with approximately the same successful rate of BOD 5 and TSS effluent level reduction. The petitioners deny these assertions.4

We consider initially whether the conclusion of the EPA that the 20-pound BOD 5 standard can be met, is arbitrary and capricious.5

1. Clinton Corn Results.

On remand, data from the Clinton Corn plant became available. Clinton Corn is a large corn wet milling plant at which a new waste treatment facility has been constructed. It utilizes much of the EPA proposed technology, including deep bed filtration. The Clinton treatment facility began operations on January 1, 1974, and data for the period from November, 1974 to September, 1975, was utilized by the EPA in setting the proposed standards.

The parties disagree as to the results achieved at Clinton.6 The EPA concludes that a monthly average of 7.1 pounds per MSBu of BOD 5 for the eleven-month period and a monthly average of 4.5 pounds for the last seven months was achieved. The petitioners conclude that the averages were 15.0 and 7.5 pounds, respectively. Even if the petitioners' higher figures are used, it appears that:

(1) the eleven-month average is less than the standard set by the EPA;

(2) the seven-month average is at least 60% below the standard set by the EPA;

(3) only one month in the last seven exceeded 50% of the standard set by the EPA; and

(4) the treatment facility removed an average of 98% of the BOD 5 in the wastewater.

Moreover, the record supports the EPA's view that lower effluent standards can be reached if all known in-plant controls to prevent spills and to equalize the effect upon the treatment facility if spills do occur are incorporated in the new plant.

The petitioners assert, however, that because only 65 5. This waste must then be treated in the regular treatment facility.

The EPA projected model plant, discussed infra, has a production capacity of 30,000 bushels and a proposed treatment facility capable of handling a wasteflow of 1.0 million gallons a day (mgd). The Clinton plant has a production capacity of 120,000 bushels or four times that of the model, but its present treatment facility is designed to treat only 3.0 mgd. As Clinton changes over to surface condensers, its treatment facility will have to be enlarged accordingly. There is no evidence in this record to suggest that this additional waste volume cannot be treated in the enlarged facility as effectively as the waste now being treated at Clinton. For our purposes, the crucial factor is that the Clinton treatment facility is effectively treating the BOD 5 wasteload it is designed to treat at a level which is consistent with the EPA proposed standards.

The petitioners also argue that the Clinton data should not be used because its treatment plant includes a "double stage" treatment process before filtration while the EPA model facility consists of a single stage. They argue that even if the Clinton plant performance for BOD 5 removal approaches that of the EPA standard, it does so only with more extensive technology at a higher cost than the EPA model facility.7

The EPA concedes that Clinton uses a double-stage pretreatment facility but properly notes that the raw waste equalization basin at Clinton is not designed to hold the raw wastes for the time required. It again calls attention to the lack of good in-plant controls.

We are convinced that the record supports the EPA's conclusion that the performance of the Clinton plant demonstrates that new plants designed and built to EPA standards can meet the BOD 5 standard of 20 pounds per MSBu.2. EPA's Model Plant.

In preparation of the standards, the EPA synthesized a model corn wet milling plant and corresponding treatment facility from data received from the corn wet milling industry and from other industries that encounter similar wastewater. The model contained most of the best technology and procedures presently used or practiced in the industry. It is from this model and its estimated ability to reduce the projected wasteloads that the EPA set the proposed new plant standards.

The petitioners challenge two of the projections made by the EPA in synthesizing its model plant. First, they assert that the EPA erred in not considering the additional wastewater flows of between 90 and 250 pounds of BOD 5 per MSBu that result from the use of wet water scrubbers. Second, they assert that the EPA erred in failing to accurately account for the higher concentrations of BOD 5 that result from the production of modified starches. These errors, they assert, render the effluent-level reductions inaccurate.

The petitioners do not dispute that the alleged additional wastewater flows could not be effectively treated by the proposed technology. They argue only that the capacity of the model treatment plant may be insufficient to handle the volumes or concentrations of this additional wastewater.

Our review of the record discloses that the petitioners' conclusion on the increased wastewater generated by the wet water scrubbers is based upon preliminary and conflicting data. The most persuasive of these inconclusive estimates is that an additional flow of 90 pounds of BOD 5 per MSBu may be created.8 As set forth above, the Clinton treatment facility has consistently achieved a 98% BOD 5 reduction over an eleven-month period during which the influent has varied considerably in terms of concentration and amount. If this additional 90 pounds were treated at the Clinton facility, the seven months BOD 5 effluent averages of 4.5 of the EPA and 7.5 of the petitioners would be increased to 6.3 and 9.3, respectively. Both of these figures remain well below the EPA proposed standard of 20 pounds of BOD 5 per MSBu.

The record also discloses that the EPA did take into consideration the increased wasteloads generated by the production of modified starches. The EPA compiled data of actual performance results from newer existing mills that produce both conventional corn syrups and modified starches and also the projections for new plants, submitted by the industry, in order to determine the wastewater flow and influent levels for the model plant. Moreover, it is undisputed that the Clinton plant produces a variety of end products including modified starches. Both the Clinton wastewater load and influent levels of BOD 5 approximate the projections set out for the EPA's model plant.

Furthermore, the data relied upon by the petitioners to show increased wasteloads because of modified starch production must be contrasted with other data showing no appreciable influent increase. As stated by a representative of CPC in a report to the EPA, there were a number of occasions when the production of modified starches increased the wasteload substantially and:

about an equal number of times when there did not seem to be any effect. No explanation has been found for this problem.

Letter from R. L. Hap of CPC International to Dr. H. E. Schwartz, Jr., June 25, 1975.

Based on the model plant data, we cannot say that the EPA ignored the effects of additional waste produced by the production of modified starches nor did it act in an arbitrary or capricious manner in calculating the effect of this waste in preparing its model plant.

3. Variability and Excursions.

The petitioners also assert that the EPA's standard for the average maximum values for thirty consecutive days of 20 pounds of BOD 5 per MSBu fails to adequately account for variability factors. They argue that if the EPA desires to set short-term thirty-day standards based on long-term averages, that the EPA must ensure that mills not violating the annual averages will also not violate the thirty-day average and daily maximums. The petitioners argue that the EPA considered variability factors only upon the expected results of its model plant and that the EPA failed to adequately consider the variability factors realized at existing plants.

The record discloses that the EPA did adequately consider variability. It took the long range effluent data from the CPC-Pekin and American Maize plants and determined the 99% to the mean monthly variability factor to be 2.7 and 1.9, respectively. This range encompasses the 2.6 variability factor for the Clinton plant estimated by the petitioners. Multiplying these factors upon the monthly average for the last seven months at Clinton as increased by additional effluent from the wet scrubbers, the range is between 12 and 17 pounds of BOD 5 per MSBu utilizing the EPA's figures and between 17.7 and 25.1 pounds utilizing the petitioners' conclusions of the Clinton data. All but the last projection is within the proposed BOD 5 standard.

Moreover, there is support for the EPA's contention that strict statistical computation of the variability factor for new plants based upon existing plant results creates an artificially high number. New plants equipped with complete in-plant controls and a properly designed and operating treatment facility should be less prone to the variations of influent created by the lack of similar facilities in existing plants and, therefore, the variation factor in a new plant should be lower.

There is also support in the record for the maximum value for any single day of 60 pounds of BOD 5 per MSBu. This single day allowance of three times the thirty-day average provides sufficient flexibility for high effluent days.

Finally, the petitioners contend that even if the variability factors utilized by the EPA are accurate and provided that the maximum day and thirty-day averages can be met, the EPA has not included sufficient consideration for the "excursions" above these levels that will occur in setting these standards. They argue that the variability factors utilized account for 99% of all predictable performance with the remaining one percent of the results from a plant expected to exceed the value calculated as the 99% maximum. The petitioners assert that because the EPA has not provided any variances for these excursions, effluent levels may exceed the allowable maximum daily level several days within a year despite the fact that the plant is meeting the yearly average.

The record reflects that this complaint is not well founded. The excursions in this industry are primarily the result of inadequate facilities, including insufficiently-sized aeration or retention basins and the failure to provide dikes and other in-plant controls to reduce the effect of in-plant spills. New plants, properly designed and constructed, should not experience this same difficulty. Dr. Raymond C. Loehr, an expert in the treatment of high strength organic wastes, has concluded:

No excursions are necessary. The variability factors that are developed are based upon existing data. There are many steps that industry can take to reduce the variability and not exceed the limitations. It is reasonable to expect that based upon the use of better treatment technology and control, attention to treatment plant maintenance and operations, and concern about in-plant waste control, industry can reduce the existing variability by at least 1%, the amount inferred by the use of the 99/50 ratio, to obtain the limitations. In fact it is reasonable to expect that such variability can be reduced by perhaps 5% by means of the variability controls identified earlier.

R. C. Loehr, EPA Memorandum on "Effluent Variability Factors," (June 26, 1975).

The record indicates that there remains the possibility of a rare occasion when the daily effluent level may exceed the maximum daily standard of 60 pounds because of inadvertent error or Act of God. The EPA has established the policy of informally working with the plant that suffers from such infrequent or isolated violations to attempt to remedy the problem causing the violation without the necessity of bringing any formal or court action. See EPA, "Guidelines for Water Pollution Enforcement," 8 (July 23, 1974). If the effluent limitations in a federally-issued permit prove to be unreasonable or cannot be met for cause beyond the control of the permittee, the permit may be revised. Id. See also, § 306(b)(1)(B), 33 U.S.C. § 1316(b)(1)(B).

The EPA has not acted arbitrarily or capriciously in its consideration of variability or excursions.

4. Conclusion.

The EPA has set the new plant effluent standards for BOD 5 at 20 pounds per MSBu as a maximum average daily value for thirty consecutive days and 60 pounds per MSBu as the daily maximum. The Clinton Corn plant treatment facility contains much of the "end of the pipe" technology proposed by the EPA. During the last seven months of available operational data, the Clinton plant achieved a thirty-day average BOD 5 effluent of 4.5 pounds, utilizing the calculations of the EPA or 7.5 pounds if the calculations of the petitioners are used. If these amounts were increased because of additional waste flow from the wet scrubbers and when a variability factor of between 1.9 and 2.7 is applied, the results range from 12 to 17 pounds of BOD 5 per MSBu utilizing the EPA's calculations and 17.7 and 25.1 pounds if the calculations of the petitioners are used.

When these final totals are considered in the light of the problems experienced in the Clinton treatment facility and the lack of in-plant controls in the Clinton processing plant, they provide adequate support for the EPA's proposed standard of 20 pounds of BOD 5 per MSBu as a thirty-day average. We find, on the basis of this record, that the BOD 5 standard set by the EPA is neither arbitrary or capricious and is in accordance with law.

We turn to a consideration of whether the conclusion of the EPA that the 10-pound TSS standard can be met is arbitrary and capricious.9

1. Clinton Corn and Other Corn Wet Milling Plant Results.

As in the BOD 5 measurements at the Clinton plant, there is a disagreement in the pounds per MSBu figure for the effluent level of TSS. The EPA concludes that a monthly average of 17.6 pounds for the last eleven-month period and an average of 16.0 pounds for the last seven months was achieved. The petitioners conclude that the averages were 34.8 and 25.6 pounds, respectively.

The EPA argues that data for the months of July, August and September of 1975 supports the TSS standard of 10 pounds per MSBu. Considering the EPA's conclusions of the Clinton data and eliminating one usually high effluent day, the following effluent levels were achieved:

(1) that the 10-pound standard would be exceeded in every month when a variability factor of between 2.2 and 2.6 is applied to these results;10 and

(2) that the TSS levels are uniformly higher than BOD 5 levels even when the treatment is achieving maximum performance levels.

The EPA suggests that this second conclusion indicates that the BOD 5 standard could be lowered. We find little merit to this suggestion. When surface condenser wastes and variability factors are given appropriate consideration, the BOD 5 standard appears to be as low as can be consistently achieved.

In each of the eleven months of data available from Clinton, the TSS effluent levels have exceeded the corresponding BOD 5 levels.11 Specifically, the TSS levels range from 3.5 to 24.5 pounds higher than BOD 5 levels. In short, the TSS standard proposed by the EPA finds no support from the Clinton plant data. The effluent levels achieved at other corn wet milling plants, namely CPC-Pekin and American Maize, is equally unsupportive.

2. Effluent Level Reduction in Other Industries.

Recognizing the lack of support for its proposed TSS standards from results within the corn wet milling industry, the EPA points to the use of deep bed filtration in other industries. It argues that these results support the proposed TSS standards for the corn wet milling industry. In our judgment, the results from other industries indicate that deep bed filtration will remove suspended solids but not with the efficiency or the consistency necessary to satisfy the proposed TSS standard for this industry.

The studies relating to municipal treatment plants indicate that deep bed filtration will remove an average of 70% of the suspended solids in the wastewater. However, the record also establishes that the removal rates range from a low of 20% to a high of 95% with no indication of consistency. Moreover, the level of suspended solid influent in most municipal treatment plants is considerably less than that found in the corn wet milling industry.

The record indicates that the strength or concentration of the influent wastewater substantially affects the success of deep bed filtration. If the wastewater has been effectively pretreated in a biological treatment system, deep bed filtration will remove a substantial percentage of the remaining suspended solids. However, one cannot say that a properly functioning system will reduce the TSS level below that of the BOD 5 level. In the winery industry, for example, treatment of heavily concentrated wastewaters similar to those found in the corn wet milling industry produce TSS levels that exceed the BOD 5 level by as much as 48 mg/1. Moreover, the deep bed filtration system there appears to remove less than 40% of the suspended solids.12

Data from the brewery and edible oil industries is largely unusable because it is set forth in milligrams per liter and because it is not clear whether deep bed filtration is used or not in these industries. This data gives some support to the finding that corn wet milling plants can meet the 1977 existing plant guidelines but little or no support that 1983 or new source standards can be met. Again, TSS levels are often higher than BOD 5 levels.

The data from Morton Frozen Foods indicates only that TSS levels are often higher than BOD 5 levels and that there are wide variations in TSS levels between the minimum and maximum days.

3. Expert Testimony.

The EPA did present testimony from an acknowledged authority on filtration of wastewaters. Dr. Robert Bauman stated that he "would expect" a new corn wet milling plant utilizing complete pretreatment facilities and filtration systems to be able to meet the proposed new source standards. However, Dr. Bauman's report is replete with qualifications and, standing alone, is insufficient grounds to support this standard, particularly in light of the inability of any existing plant or pilot project to demonstrate the efficacy of this technology to consistently meet the proposed standard.

4. Conclusion.

The decision of the EPA that new corn wet milling plants can meet a TSS standard of 10 pounds per MSBu is an arbitrary and capricious one.

There remains the question of what our response to this inadequacy should be. Ordinarily, we would simply remand and instruct the EPA to develop a new standard for TSS and retain jurisdiction to review that new standard in an accelerated appellate proceeding. But there has been one remand, and more than eighteen months have been lost in achieving the national goal of cleansing our water. We, thus, feel the need to adopt a more flexible approach, one which will, hopefully, permit a new standard to be implemented forthwith. We are convinced from this record that a standard of 25 pounds of TSS per MSBu would not be an arbitrary and capricious one. Results at the Clinton Corn plant lend support to such a standard.13

On remand, therefore, the EPA may adopt a new source standard of a maximum average of daily effluent levels for thirty consecutive days of 20 pounds of BOD 5 and 25 pounds of TSS per MSBu and a maximum daily effluent level of 60 pounds of BOD 5 and 75 pounds of TSS per MSBu. We urge it to do so.

The EPA retains the authority, however, to take additional evidence in support of a lower TSS standard. If it adopts that course, we retain jurisdiction. We direct that the administrative process be completed in sixty days. The only question to be addressed on that limited remand would be the TSS standard. Issues raised and answered in CPC I or in this opinion should not be relitigated.

III. Costs.

Section 306(b)(1)(B), 33 U.S.C. § 1316(b)(1)(B) states in part:

In establishing or revising Federal standards of performance for new sources under this section, the Administrator shall take into consideration the cost of achieving such effluent reduction, and any non-water quality environmental impact and energy requirements.

In CPC I, we instructed the EPA to reconsider its projections on the cost of achieving the effluent reduction proposed in its standards. We asked that operating and capital costs be projected independently and that the cost computations be based on current economic data. We noted that costs were of particular importance to this industry because profit margins are usually low, raw material costs are volatile and competitive substitutes sometimes make it difficult to pass on increased costs to the consumer. CPC International Inc. v. Train, supra at 1050-1051.

The EPA states that it has responded to the instructions of this Court. It asserts that its obligation under the Act is to demonstrate that the costs to be incurred in meeting the new source standard are reasonable. It asserts that capital and operating costs meet this standard.

The petitioners assert that the Act requires a "cost-benefit analysis" and that the EPA failed to make such an analysis; that the technology proposed by the EPA to meet the new source standards is insufficient to achieve the effluent reduction necessary; that the capital and operating costs for the necessary technology will impose an unreasonable burden on the industry; and that calculation errors make the economic analysis completed by the EPA inadequate.

A. A Cost-Benefit Analysis is Not Required; the Appropriate Test is One of Reasonableness.

The 1977 existing plant guidelines are to be determined by the use of the "best practicable control technology currently available." § 301(b)(1)(A), 33 U.S.C. § 1311(b)(1)(A). The legislative history of the Act indicates that the term "practicable" is to limit the use of available technology only where the additional technology necessary to achieve a marginal level of effluent reduction is wholly out of proportion to the cost realized. A Legislative History of the Water Pollution Control Act Amendments of 1972, 93rd Cong., 1st Sess. (Comm.Print 1973), at 170 (hereinafter cited as Leg.Hist.); American Iron and Steel Institute v. E. P. A., 526 F.2d 1027, 1051 (3rd Cir. 1975).14 The 1983 existing plant guidelines, set by the application of the "best available technology economically achievable," § 301(b)(2)(A), 33 U.S.C. § 1311(b)(2)(A), are to be governed by a standard of reasonableness without the necessity of a thorough cost-benefit analysis. Leg.Hist. at 170; Amer. Meat Institute v. Environ. Protect. Agcy., 526 F.2d 442, 462-463 (7th Cir. 1975).

The new source standards are to attain the "greatest degree of effluent reduction" achievable through the use of the "best available demonstrated control technology." Section 306(a)(1), 33 U.S.C. § 1316(a)(1). This standard exceeds the effluent level reduction of that required in the 1977 existing plant guidelines for the corn wet milling industry while the proposed 1983 existing plant guidelines are the same as the new source standards. The legislative history indicates that Congress felt that new plants could realize effluent level reductions at a lower cost than existing sources because of the lower cost of constructing a new plant in comparison to retrofitting an existing facility and because technical alternatives may be made available for new plants which would not be suited for existing plants. Leg.Hist. at 172 and 797.

There is no language in § 306 requiring a cost-benefit analysis. Rather, the EPA is required only to take costs under "consideration." We conclude, therefore, that a cost-benefit analysis is not required in determining the reasonableness of the cost of achieving the new source standards. Accord, American Iron and Steel Institute v. E. P. A.,supra at 1059. What is required for new source standards is a thorough study of initial and annual costs and an affirmative conclusion that these costs can be reasonably borne by the industry.

B. The Technology Proposed by the EPA is Sufficient; the EPA's Capital Cost Estimates are Accurate.

The petitioners' argument that in-plant facilities, in addition to those proposed by the EPA, are necessary to meet the new source standards does not have support in this record. They suggest that additional evaporators costing $4.2 million must be installed solely for the purpose of reducing the volume of wastewater and to prevent overloading of the treatment facility. The record, however, supports the EPA's contention that evaporation is a common production process used in the industry without regard to pollution controls. Evaporation is used in most corn wet milling plants to concentrate the liquid corn syrup into products which can be readily shipped without the need to transport excess water. This process is also used to concentrate steepwater to recover the solids in the water for animal feed. Indeed, the EPA model new plant projects an estimated capital cost of $3 million for evaporation equipment in the production process. The only support for petitioners' claim that additional evaporation equipment is required is their own conclusory statement.

The petitioners' contention that additional entrainment separators, costing $634,000, are necessary must also be rejected. They have failed to support the need for this technology. Moreover, they argue, in another section of their brief, that the effectiveness of such separators has not been documented.

The petitioners also argue that the EPA's estimate of the incremental cost of installing surface condensers of $250,000 is too low. They claim that this cost will be $1 million. They argue that a new 30,000-bushel plant will require six condensers rather than the three proposed by the EPA and that the cost of each condenser is twice that projected by the EPA.

The petitioners base their conclusion that six condensers are needed on the experience of the CPC-Pekin plant. In our view, the comparison is not appropriate. The CPC-Pekin plant has a capacity of at least 60,000 bushels and a present water flow much larger than the amount projected for the proposed model plant. There is also no support for the petitioners' contention that each surface condenser will cost twice the amount projected by the EPA.

We also reject petitioners' contention that the capital costs of the treatment facility will significantly exceed those estimated by the EPA. All parties agree that the basic biological facility, designed to treat a BOD 5 load of 400 pounds per MSBu, would be approximately $3 million.

There is disagreement as to the cost of constructing the filtration unit. The petitioners' contention that the cost would be about $100,000 more than that estimated by the EPA15 has strong support in the record.16

In summary, the record supports the conclusion of the EPA that additional in-plant equipment, surface condensers, would have a capital cost of $250,000. The record also supports the conclusion that a basic biological treatment facility capable of handling 400 pounds of BOD 5 per MSBu would have a capital cost of $3 million. An additional capital cost of $260,000 would be realized in purchasing and installing the equipment needed for the filtration system.17 The record does not support the contentions of the petitioners that additional equipment is necessary to meet the proposed new standards.

C. The EPA's Revised Operating Cost Estimates are Not Arbitrary or Capricious.

The operating cost estimates include capital recovery costs and operating and maintenance expense. The EPA estimates an annual capital recovery cost of $402,000 and an annual operating expense of $300,000. The petitioners suggest that annual capital recovery costs will be $1,359,000 and annual operating expense will be $906,000.

The major differences in capital recovery expense relates to petitioners' contentions that technology in addition to that suggested by the EPA is required and that the capital cost of some of the equipment is underestimated by the EPA. We have answered those contentions in an earlier section of this opinion. On the basis of our earlier discussions and assuming an annual capital recovery rate of approximately 12% (a rate agreed to by both parties), the annual capital recovery costs would be approximately $414,000.

The parties agree that the operation and maintenance expense for the surface condensers will be $40,000. The EPA projects that similar expenses for the treatment facility, including filtration, will be $260,000. The petitioners project that these expenses will be $414,000 annually. The EPA's estimates are based upon the actual operating expense at the CPC-Corpus Christi plant from January to May, 1974, while the petitioners' estimate is based on data from the CPC-Pekin plant during this same time period. Both of these plants employed some of the elements of the EPA's proposed treatment facility, but neither utilized a filtration system. Both plants experienced higher operational costs than would be realized at a new plant because the treatment facilities were added in a piecemeal fashion and have been subject to numerous malfunctions. The wastewater flow to the Pekin treatment facility is 1.0 mgd, the amount proposed for the model plant, while that at Corpus Christi is less.

While we think petitioners' estimates to be the more accurate, we cannot say that EPA's estimates are arbitrary and capricious. We, accordingly, accept EPA's finding that annual operating and maintenance will be approximately $300,000.18

In conclusion, the record supports the projection that the total operating cost for water pollution control equipment in a new 30,000-bushel corn wet milling plant will be approximately $714,000, including $414,000 capital recovery expense and $300,000 for actual operating and maintenance expense.

D. Despite Calculation Errors by the EPA, the Cost of Achieving the Effluent Reductions Required by the New Source Standards is Not Unreasonable.

To determine the economic impact of meeting the new source standards, the EPA hired a consultant to prepare an economic analysis. Utilizing the treatment facility proposed by the EPA and historical cost data made available by the industry, the consultant estimated that a new 30,000-bushel plant would realize a return of investment of 17.1% without controls and 15.8% with the proposed controls.19 He also forecasted a return on sales of 14.1% without controls and 13.4% with controls. He estimated larger returns on investment and sales for 60,000 and 90,000-bushel plants. Based on his analysis, the EPA concluded that the costs incurred were "reasonable" and set forth the proposed new standards for comment by interested parties as required by law.

More than thirty days after the formal comment period expired, a representative of the corn wet milling industry informed the EPA that the consultant had erred in determining revenue utilizing a dry basis corn syrup price with an estimation of the amount of product produced on a liquid basis. The correct calculation requires a liquid corn syrup price. The EPA's consultant, after review, conceded that he had erred. He admitted this error was substantial and suggested that estimated revenues should be reduced by as much as 20% and after-tax income by as much as 50%.20

This error has a major impact on the earlier projections. A return on investment with this revision is now estimated to be 6.9% without controls and about 5.8% with controls. The return on sales with this revision is now calculated to be 6.6% without and about 5.7% with controls.21

Despite this reduction of the estimated economic viability of a new corn wet milling plant with or without proposed controls, the consultant concluded that all three-sized plants remain "economically viable in relation to general food processing industry standards of performance." Neither the consultant nor the EPA elaborated upon this conclusion.

The petitioners have not aided the Court in this matter. After raising the calculation error, they do not specifically challenge, either in their briefs or at oral argument, the conclusion of viability by the EPA and its consultant. They state only that:

(a)ny further sensitivity adjustment to reflect lower revenues, higher waste treatment plant costs, or higher costs of capital, could reduce the net present value to the point at which new corn wet milling plants were not economically viable.

The only implication that can be drawn from this failure is that the rate of 6.9% on invested capital is adequate, a conclusion that we would find hard to accept or that the actual rate of return might be higher than that and it was better to let sleeping dogs lie.22 No argument is made that the relatively minor sensitivity adjustment which we have made with respect to additional filtration capital cost of $100,000 is sufficient to affect the economic viability of a new plant.

Moreover, no argument is made that new plants without controls are economically viable with an estimated after-tax return on invested capital of 6.9% but are not viable if the additional cost of controls reduces the return on invested capital to approximately 5.7%.

Based on a careful review of a generally unsatisfactory record, we reluctantly conclude that the EPA has not abused its discretion or acted arbitrarily in ascertaining that the cost of achieving the proposed effluent level reduction for new sources in this industry is reasonable.

V. Summary.

We hold that the new source standard of 20 pounds of BOD 5 per MSBu as a maximum average of daily effluent values for a thirty consecutive day period set forth by the EPA is not arbitrary or capricious.

We hold that the new source standard of 10 pounds of TSS per MSBu as a maximum average of daily effluent values for a thirty consecutive day period set forth by the EPA is arbitrary and capricious.

We hold that neither capital costs nor capital recovery and operating costs have been shown to be arbitrary and capricious.

We remand to the EPA with directions to it to revise the TSS standard to 25 pounds per MSBu or to compile additional evidence to support a lower standard. If the EPA decides to present a lower TSS standard, we retain jurisdiction and direct that the administrative process be completed within sixty days. The only question to be addressed on that limited remand would be the TSS standard. Issues raised and answered in CPC I, or in this opinion, should not be raised.

Costs shall be divided equally among the parties.