On rehearing, counsel for the plaintiff have expressed great concern lest the tax laws be misapplied by an adherence to our former opinion. However, accepting as we must, the application of the law to unincorporated wholly-controlled branches or businesses located in other jurisdictions as set forth in Butler Brothers v. McColgan, 17 Cal.2d 664 [ 111 P.2d 334 ], 315 U.S. 501 [ 62 S.Ct. 701 , 86 L.Ed. 991 ], the conclusion is irresistible that the same rule should apply to incorporated wholly-controlled branches or businesses so located. This is demonstrated by the following portions of our former opinion dealing with that subject.

The plaintiff, a California corporation, brought these two actions to recover franchise taxes paid under protest for the taxable years 1937 and 1938. The trial court rendered judgment for the plaintiff in both actions and the defendant appealed. The appeals are presented on a single set of briefs.

Prior to 1929, Edison Brothers engaged as partners in the retail shoe business in Atlanta, Georgia. They, expanded their operations to include stores in Louisiana, Alabama and Tennessee. In 1926, the business was organized as a corporation under the laws of the state of Georgia. In 1929, a corporation, Edison Brothers Stores, Inc., was organized under the laws of Delaware and acquired all of the capital stock of the Georgia corporation. The Delaware corporation established and operated stores in several other states, including Missouri, Illinois, Ohio, Tennessee and California. In 1930, the stores in Missouri and Tennessee were incorporated under the laws of those states. An Illinois corporation was also organized. In 1932, the plaintiff was incorporated under the laws of this state. In all, fifteen subsidiary corporations were formed. They are located in as many states and their capital stock is owned and controlled by the Delaware corporation. Members of the Edison family, who reside in St. Louis, Missouri, constitute the officers and directors of all the subsidiary corporations and of the Delaware, or parent, corporation. A central management division, a central purchasing department, a central distributing department, a central store operations department, a central advertising department, and various other central administrative departments function in St. Louis. These centralized departments determine operating policies and keep the main accounting records for all of the subsidiaries. The parent corporation manufactures no goods, but purchases ladies’ shoes, shoe findings, accessories, hand bags, hosiery and costume jewelry for distribution and retail sale by itself and each of its subsidiaries. The quantity of merchandise for shipment to each store for resale is determined in St. Louis from a statistical analysis of daily reports sent in by the store managers. The foregoing services are performed by the Delaware corporation pursuant to agreements between it and each subsidiary by which the cost of merchandise plus a specified percentage is charged to each subsidiary, in addition to a proportion of the general overhead.

The plaintiff sells the merchandise received by it from the parent corporation exclusively within the state. It conducts no interstate sales. For each of the taxable years 1937 and 1938, it filed a franchise tax return showing gross receipts less claimed deductible expenses in accordance with a separate accounting system based on the agreement between the subsidiary and the parent corporation. For the taxable year 1937, the plaintiff showed a net (1936) income of $68,-299.90, and a tax computed thereon of $2,731.99, which it paid. Subsequently the plaintiff received a notice of additional assessment in the sum of $12,329.73. After protest and oral hearing the additional assessment was recomputed and redetermined at $5,161.09 with interest thereon, which the plaintiff paid under protest, and thereupon filed a claim for refund which was denied. The return for the taxable year 1938, reported a net (1937) income of $17,808.58 (although there was an increase of $900,000 in sales) as to which the plaintiff paid a franchise tax of $712.34. Similarly the franchise tax commissioner assessed an additional tax of $13,789.70, which after hearing on the plaintiff’s protest was reduced to $6,313.35. This sum with interest the plaintiff paid under protest and filed a claim for refund which was rejected.

The additional assessments resulted from the commissioner’s application of a three-factor formula to allocate the net income of the plaintiff, considered as an integral unit in an entire business comprising the parent and all of its subsidiary corporations. The commissioner, with reliance on Butler Brothers v. McColgan, supra, 17 Cal.2d 664 , 315 U.S. 501 , contended that on the basis of the asserted unitary relationship, the figures submitted reflected the true net income attributable to the business done within this state.

By a. finding that the plaintiff transacted only intrastate business the trial court drew the conclusion that the separate accounting method, as distinguished from the method based on formula allocation, properly reflected the net income attributable to business done within the state. The court also found that the plaintiff’s book entries were accurate and reasonable. Implicit in the findings is the conclusion that there was no falsification or bad faith in the manner of bookkeeping employed by the plaintiff, nor in the entry of any of the items contained therein, nor any intent thereby to reflect an improper income or create an artificial loss. The judg ments may therefore be proper if the fact of the organization of the California business as a domestic corporation distinguishes it in principle from the case of Butler Brothers v. McColgan, supra ( 17 Cal.2d 664 , 315 U.S. 501 ).

Persons may adopt any lawful means for the lessening of the burden of taxes which in one form or another may be laid upon properties or profits. (Pioneer Express Co. v. Riley, 208 Cal. 677, 687 [ 284 P. 663 ].) It was also reiterated in that ease that courts, in interpreting statutes levying taxes, may not extend their provisions, by implication, beyond the clear import of the language used, nor enlarge upon their operation so as to embrace matters not specifically included. In ease of doubt, construction is to favor the taxpayer rather than.the government. But unquestionably the Legislature may enact measures to prevent avoidance of taxes otherwise properly laid.

In this state the Legislature has enacted the Bank and Corporation Franchise Tax Act (Stats. 1929, p. 19, as amended; 3 Deering’s Gen. Laws, Act 8488), herein referred to as the act, pursuant to the power vested in it by section 16 of article XIII of the state Constitution. By section 4 of the act a corporation doing business within the limits of the state, for the privilege of exercising its corporate franchise in the state, is subject to an annual tax measured by four per cent of its net income earned during the preceding fiscal or calendar year, the minimum tax being $25.

Section 10 of the act, in effect during the period involved, provided that if the entire business of the corporation is not done within the state, the tax should be measured by that portion thereof derived from business done within the state, determined by a method of allocation upon the basis of sales, purchases, expenses of manufacture, payroll, value and situs of tangible property, or any of them or of other factors, or by such other method of allocation as may be fairly calculated to assign to the state the portion of net income reasonably attributable to business done within the state.

Section 12 provided that the net income should be computed in accordance with the method of accounting regularly employed by the taxpayer, but if that method does not clearly reflect the income, “the computation shall be made in accordance with such method as in the opinion of said commissioner does clearly reflect the income.”

Section 14, in effect during the period involved, dealt with two or more corporations owned or controlled directly or ii> directly by the same interests, with authority in the commissioner to apportion or allocate the income or deductions among them if he deemed it necessary to prevent evasion of taxes, or to require a consolidated report or such other facts as might be deemed essential to determine the amount of net income.

Butler Brothers v. McColgan, supra ( 17 Cal.2d 664 ; 315 U.S. 501 ), involved the application of section 10 of the act to determine the net income attributable to business done within this state by a foreign corporation exercising its corporate franchise here. Butler Brothers was an Illinois corporation engaged in purchasing dry goods and general merchandise from manufacturers and distributing them to wholesale houses owned by it in seven states, including California, for resale to retailers only. It maintained a central buying division. The cost of goods and the transportation charges were entered on the books of the house which received the goods. The actual cost of operating the central buying division was allocated among the houses, as also were common expenses, such as executive salaries, accounting and advertising. No question was raised as to the accuracy of the charges to each house. By the method of computation employed by it the taxpayer reported a loss of $82,851 for the year 1935, from operations in California. But in that year operations of all houses yielded a net profit of $1,149,677. The commissioner, pursuant to section 10 of the act, by a method of formula allocation—that is by averaging the percentages which (1) the value of real and tangible personal property, (2) wages, salaries, commissions and other compensation paid to employees, and (3) gross sales, less returns and allowances, attributable to the California house, bore to the corresponding items of all houses—allocated to California 8.1372 per cent of the entire net profit. Instead of a loss of $82,851 for operations in California, this method attributed to California business a net profit, of $93,500. The facts also showed that of the gross sales of $66,326,000 by all the houses, $5,206,000 thereof were by the California house, and that the gross volume of sales in California was substantially in the same proportion to the entire gross business as the percentages resulting from the formula allocation method. The use of the formula method of allocating the amount of business attributable to operations in California was upheld by this court and by the Supreme Court of the United States. This court said ( 17 Cal.2d at p. 667-668 ) : "It is only if its business within this state is truly separate and distinct from its business without this state, so that the segregation of income may be made clearly and accurately, that the separate accounting method may properly be used. Where, however, interstate operations are carried on and that portion of the corporation’s business done within the state cannot be clearly segregated from that done outside the state, the unit rule of assessment is employed as a device for allocating to the state for taxation its fair share of the taxable values of the taxpayer.” This court pointed out that the general test for the unit rule of assessment was the unity of use and management; that is, if the operations in California contributed to the net income derived from the entire operations in the United States, then the entire business is so clearly unitary as to require a fair system of apportionment by the formula method in order to prevent either overtaxation or undertaxation. It was also observed that the purchasing and the sales activities were interdependent and supplemented each other to form one single, unitary business; that purchase price differentials were affected by the unitary power; and that sales in each of the states must be proportionately credited with the profits resulting from the purchasing power which such sales in the aggregate produced. Other factors in addition to purchasing which indicate the unitary character of the business were pointed out, namely, the central advertising, accounting and management divisions. The employment of the allocation method in any ease of unitary business was approved so long as it was not intrinsically arbitrary and produced no unreasonable result; also that where a business is nnitary in character, so that its separate units could not fairly be considered by themselves and the whole business in the several states derived a value from the unity of use, allocation of income upon a reasonable formula was sustainable. This court declared that the burden was upon the taxpayer to show that the method was arbitrary or produced unreasonable results. It was considered that the unitary nature of the business was definitely established by (1) unity of ownership ; (2) unity of operation evidenced by central purchasing, advertising, accounting and management; and (3) unity of use in the centralized executive force and general system of operation. It was concluded that the application of the formula to allocate the proportion attributable to this state of net income of an entire unitary system was within the purview of the statute which measured the tax on the income derived from business done within the state.

At the time of the Butler Brothers decision the latest expression of the Supreme Court of the United States upholding the propriety of the use of formula allocation where unitary management was shown, was said to be in 1936 (North American Cement Corp. v. Graves, 269 N.Y. 507 [ 199 N.E. 510 ], 299 U.S. 517 [ 57 S.Ct. 311 , 81 L.Ed. 381 ]). In 1941 that court affirmed the decision of this court in the Butler Brothers case with an opinion upholding the formula method of allocation in such cases. It was held that one who attacks a formula of apportionment assumes the burden of showing that it has resulted in the taxation of extra-territorial values. It was there declared that since all the factors in the conduct of the Butler Brothers business were essential to the realization of profits, the business was properly treated as unitary; that the apportionment formula based on property, payroll and sales was not inappropriate where the contribution of the California branch to the advantages of centralized management of the unitary enterprise and to the entire net income earned was not questioned; and that the employment of the formula method was fairly calculated to assign to California that portion of the net income reasonably attributable to business done within the state.

The Butler Brothers case demonstrates that the mere production of a separate accounting system for one of the branches of an entire or unitary business is insufficient without more to overthrow an apportionment formula based on the appropriate factors of sales, property and payroll. The fairness of the three-factor formula was thereby settled. Only if the taxpayer can meet the burden of showing by sufficient evidence that the allocation formula does not apportion to California that part of the entire net income reasonably attributable to the business done in the state will the court disapprove its use in the particular case. The evidence introduced to overcome the formula method of allocation must go to each element of the formula equation. (Butler Brothers v. McColgan, supra, 17 Cal.2d at p. 677-678 .)

In the present ease all of the elements of a unitary business are present—unity of ownership, unity of operation by centralized purchasing, management, advertising and ac counting, and unity of use in the centralized executive force and general system of operation. The business of the parent and all of its subsidiaries is owned and managed under one centralized system, to the same extent as in the Butler Brothers case and other eases considered therein. Thus the business is unitary regardless of the fact that in the Butler Brothers case there was but one corporation involved, owning as parts of the unitary system seven different branches in as many states, and that in the present case there is a parent corporation owning and controlling as units of one system fifteen different branches organized as corporations in as many states. No difference in principle is discernable. If the crux of the matter is to ascertain that portion of the business which is done within this state, then the same considerations justify the use of the formula allocation method in the one case as in the other.

The plaintiff contends that the statute does not authorize the commissioner to apply the apportionment formula when the taxpayer is a California corporation, and is not the foreign (parent) corporation as in the Butler Brothers case.

It may be assumed that section 14 of the act, which authorizes the commissioner to require a consolidated return, contemplates two or more corporations, both or all of which are taxable as doing business within the state. (See Bay Cities Tramp. Co. v. Johnson, 8 Cal.2d 706 [ 68 P.2d 710 ].) Power to apply the formula allocation in this or in the Butler Brothers case is not derived from the authority to require the filing of consolidated returns, since the latter indicates that the income of the group will be taxed as a unit. The power flows from the authorized method of ascertaining the income attributable to a taxpayer’s activities within the state; and by a parity of reasoning the authority to pursue the method is present whenever activities are partially within and partially without the state (section 10), as in the case of a unitary system, whether the integral parts of the system are or are not separately incorporated; or (under section 12) the accounting system of the taxpayer does not dearly reflect the income, which may be the case when it is part of a unitary system. The Legislature by enacting the foregoing sections, and the courts in the Butler Brothers decisions, recognized that the separate accounting method is appropriate to determine the true income of a separate business; but that when the business is not separate, and is an integral part of a larger and unitary system, the separate accounting is inadequate and unsatisfactory in ascertaining the true result of the activities and values attributable to that business. If the operation of the portion of the business done within the state is dependent upon or contributes to the operation of the business without the state, the operations are unitary; otherwise, if there is no such dependency, the business within the state may be considered to be separate. Within that test it may not properly be contended that the taxpayer in the present ease is doing a separate business.

The ascertainment of income by the apportionment method is not necessarily a disregard of the corporate entity nor an extension of the provisions of the statute by implication. Formula allocation is merely a method of ascertaining the true income attributable to the plaintiff’s business, and is a method of ascertainment authorized by the statute. In In re Morton Salt Co., 150 Kan. 650 [ 95 P.2d 335 at pp. 338-339 ], the propriety of disregarding the corporate form by the allocation method of ascertaining income in a unitary business was indicated, with citation of other cases. By a memorandum opinion in People ex rel. Federal Motor Truck Co. v. Lynch, 264 N.Y. 679 [ 191 N.E. 623 ] (cert. den. 296 U.S. 588 [ 56 S.Ct. 99 , 80 L.Ed. 416 ]), the New York Court of Appeals upheld the determination of the state tax commissioner modifying an assessment against a subsidiary corporation upon the basis of the combined net income and combined segregated assets of the parent, a Michigan corporation, and its subsidiaries of which only the relator had property or did business in New York State.

Reliance on Burroughs Adding Machine Co. v. Tax Com., 237 Wis. 423 [ 297 N.W. 574 ], does not avail the plaintiff. (See, also, Northern States Power Co. v. Tax Com., 237 Wis. 433 [ 297 N.W. 578 ].) The court in that case rejected a particular ratio method of ascertaining the income of the taxpayer as unauthorized by the statute. There the taxpayer was a subsidiary corporation of a foreign corporation parent. The court conceded the fairness of an appropriate authorized allocation method to reach the desired objective, that is, to ascertain the amount of income in the state of the taxpayer as a member of a unitary system, whether the system was. operated by one corporation, or by a parent and subsidiary-corporations.

Cases, such as Hans Rees’ Sons v. State of North Carolina, 283 U.S. 123 [ 51 S.Ct. 385 , 75 L.Ed. 879 ], where the court refused to approve the result reached by the formula allocation method because the part of the taxpayer’s entire income apportioned to the state was not reasonably attributable to the processes conducted within its borders, are in harmony with the foregoing conclusions.

In any event, since the business transacted by the parent and its subsidiaries in the present case was a unitary business, the accident of different labels which may have attached to the various subsidiaries does not change the result. The factors by which the local branch contributed to the realization of profits to the entire business in the Butler Brothers case are likewise present in this case.

The plaintiff contends that nevertheless the application of the formula allocation produced an arbitrary and unreasonable result. The contention is based on the assertions that there was no proof by the commissioner that the formula method was proper or truly reflected the taxpayer’s income; that there was no issue of tax evasion; and that, as the court found, the accounts of the plaintiff were accurate and reasonable. The court made no express finding on the reasonableness of the application of the formula method employed by the commissioner.

The evidence on the issue of the propriety of the formula allocation method to reflect the plaintiff’s income was sufficient when the unitary method of ownership, use and operation of the business of the parent and its subsidiaries was established. It was then incumbent upon the plaintiff to produce evidence to meet the burden devolving upon it to show that the formula allocation method was arbitrary and unreasonable. The plaintiff’s evidence, however, consisted solely of the presentation of its separate accounting, and the accuracy and reasonableness of the entries thereof. That evidence did not overcome the correctness of each element of the formula equation, nor the reasonableness of the application thereof to ascertain net income in this case. Tax evasion is not a factor which must be present before the formula allocation method may be employed. Reliance upon the accuracy and reasonableness of the separate accounting and the entries thereof is merely an argument for the application of a different formula, namely, a “fair price” standard. The plaintiff does not establish the unreasonableness of the formula alloca tion method by showing the reasonableness of its book entries. Its burden is to establish affirmatively that the formula equation produces an arbitrary and unreasonable result. The plaintiff makes no contention that the percentage of the entire net income allocated to this state by the formula equation had no reasonable relation to the proportion which the total sales in California bore to the total sales of the unitary business. Its position in this respect is explainable, inasmuch as the California sales for each of the years involved amounted to more than 20 per cent of the total sales of the unitary business for that year, which was a higher percentage in each instance than that allocated to California by the commissioner. There is no necessary inconsistency between the accuracy and fairness of the taxpayer’s accounting and the different result obtained by the formula method of allocating income. For taxation purposes the one does not impeach the other. (Butler Brothers v. McColgan, 315 U.S. at p. 507 [ 62 S.Ct. 701 , 86 L.Ed. 991 ].) There is therefore no evidence to support a finding, had one been made, of unreasonableness in the employment of the formula allocation method to ascertain the plaintiff’s share of the income of the unitary business. On the contrary, it affirmatively appears that the result from the method pursued was just and reasonable.

The Supreme Court of the United States has answered the plaintiff’s constitutional objections by its holding in the Butler Brothers case ( 315 U.S. at p. 507 ; see, also, Great Atlantic & Pac. Tea Co. v. Grosjean, 301 U.S. 412 [ 57 S.Ct. 772 , 81 L.Ed. 1193 , 112 A.L.R. 293 ]), that to compute the income allocable to one state of a unitary business operated in several states by a reasonable formula method is not to tax extraterritorial values.

That the authority existed to ascertain the net income of the plaintiff on a formula allocation basis in the absence, as here, of any equity which would prevent it, is made manifest by the foregoing observations. The trial court therefore incorrectly concluded that the plaintiff was immune from the appropriate exercise of that authority.

It is also contended by the plaintiff that the notice of the additional tax for the taxable year 1937 was not mailed within the prescribed time. The notice was dated November 22, 1940, and" it may be assumed was mailed on that date. Section 25 of the act as amended in 1933, provided that notice of additional tax must be mailed within three years after the return was filed. The return for the taxable year 1937 was filed March 15, 1937. According to the computation of the time limitation as enacted in 1933, the notice should have been mailed before March 15, 1940, and having been mailed on November 22, 1940, was mailed late. However, in 1939, the Legislature again amended section 25 to provide that the notice should be mailed within four years after the return was filed. This extension became effective before the expiration of the former three-year period for mailing notice of additional tax for the 1937 taxable year. Section 21 of the 1939 amendatory act provided that, since the act prescribed a tax levy for the usual current expenses of the state, it should take effect immediately and be applied in the computation of taxes accruing subsequent to December 31, 1938. In invoking the three-year rather than the four-year statute, the plaintiff makes the same contentions that were disposed of in the case of Mudd v. McColgan, ante, p. 463 [ 183 P.2d 10 ], this day decided. That ease involved the Personal Income Tax Act of 1935. A 1939 amendment (Stats. 1939, p. 2528) increased from three to four years the time for mailing a notice of a deficiency tax. Section 23 of the amendatory act contained the same provision as section 21 of the 1939 act amendatory of the Bank and Corporation Franchise Tax Act. In that case likewise the amendatory act became effective prior to the expiration of the three-year period, and the notice of deficiency tax was mailed after the three-year but within the four-year period. We held that the four-year statute was applicable. What was said there affords a complete answer to the plaintiff’s arguments here, and the result in that case is controlling. The additional tax in this case, as the deficiency tax in the case of Mudd v. McColgan, was a continuing liability and the Legislature had the power before the expiration of the prior period to extend the time within which enforcement could be commenced in the absence, as here, of an express declaration to the contrary.

It follows that the assessment of the additional tax in each case was proper.

The judgments are reversed.

Gibson, C. J., Edmonds, J., Carter, J., Schauer, J., Spence, J., and Peek, J. pro tern., concurred.