CHAPTER I. The integral in an abstract space. $1. Introduetion. Apart from funetions having as argument a variable number, or system of » numbers (point in n-dimensional space), we shall discuss in this book functions for which the inde- pendent variable is a set of points, Functions of this kind have occurred already in classical Analysis, in several important particular cases. But they only. began to be studied in their full generality during the growth of the Theory of Sets, and_ in close relation to the parts of Analysis direetly based on that thec Tf we are, for instanes, given a function f(a) integrable on every interval, then by associating with each interval Z the value of the integral of f(x) over L, we obtain a function #(J) that is a function of an interval. Similarly, by taking multiple integrals of functions Hay yy %) Of n variables, we are led fo consider functions of more general sets ying in spaces of several dimensions, the argu- ment [ of our function F(J) being now replaced by any set for which the integral of our given function fliry wy, .., @) is defined. We dwell on these examples in order to emphasize the natural connection between the notion of integral (in any sense) and that of function of a set. Needless, to say, there are many other examples of functions of a set. Thus in elementary geometry, we have for instance, the length of a segment or the area of a polygon. The class of values of the argument of these two functions (the length and the area) is in the first case, the class of segments and in the second, that of polygons. ‘The problem of extending these classes gave birth to the general theories of measure, in which the notions of length, area, and volume, defined in elementary geometry for fa restricted number of figures, are now extended to sets of points 1 S Saks, Theory of the Intezeal2 CHAPTER 1. ‘the integral in an abstract space. of much greater diversity. It is, nevertheless, remarkable that these researches arose far less from problems of Geometry than from their connection with problems of Analysis, above all with the tendency to generatize, and to render more precise, the notion of definite integral. This connection has occasionally found expression even in the terminology. Thus du Bois-Reymond called integrable the sots that to-day are said to be of measure zero in the Jordan sense. The theories of measure have, in the course of their development, been modified in accordance with the changing requirements of the Theory of Functions. In our account, the most important part will be played by the theory of H. Lebesgue. Lebesgue’s theory of measure has made it possible to dis- tinguish in Euclidean spaces a vast class of sets, called measurable, in which measure has the property of complete additivity — by this we mean that the measure of the sum of a sequence, even infinite, of measurable sets, no two of which have points in common, is equal to the sum of the measures of these sets. The importance of this class of sets is due to the fact that it includes, in particular, (with their classical measures), all the sets of points occurring in problems of classical Analysis, and further, that the fundamental operations applied to measurable sets lead always to measurable sots. It is nevertheless to he observed that the ground was prepared for Lebesgue’s theory of measure by earlier theories associated with the names of Cantor, Stolz, Harnack, du Bois-Reymond, Peano, Jordan, Borel, and others. These earlier theories have, however, to-day little more than historical value. ‘They, too, were suitable instruments for studying and generalizing the notion of integral understood in the classical sense of Riemann, but their results in this direction have been largely artificial and accidental. It is only Lebesgue’s theory of measure that makes a decisive step in the development of the notion of integral. This is the more re- markable in that the definition of Lebesgue apparently requires only a very small modification of a formal kind in the definition of integral due to Riemann. To fix the ideas, let us consider a bounded function /(#, y) of two variables, or what comes to the same thing, a bounded fune- tion of a variable point defined on a square Ky. In order to deter- mine its Riemann integral, or more precisely, its lower Riemann- Darboux integral over K,, we proceed as follows. We divide thets 11 Introduction. 3 square Ky into an arbitrary finite number of non-overlapping rectangles Ry, Ry, .... R,, and we form the sum ay Svj-m(Ri) A where »; denotes the lower bound of tho function / on ki, and m(2,) denotes the area of K;. The upper bound of all sums of this form is, by definition, the lower Riemann-Darbour integral of the function f over Ky. We define similarly the upper integral of f over Ky. Tf these two extreme integrals are equal, their common value is called the definite Riemann integral of the function f over K,, and the function f is said to be integrable in the Riemann sense over Ky. ‘The extension of measure to all sets measurable in the Lebesgue sonse, has rendered necessary a modification of the process of Rie- mann-Darboux, it being natural to consider sums of the form (1.1) for which {Rj}/-1.2,...0 i8 a subdivision of the square Ky into a finite number of arbitrary measurable sets, not necessarily either rectangles or elementary geometrical figures. Accordingly, m(R,) is to be un- derstood to’ mean the measure of Ry. The v retain their former meaning, i.e. represent the lower bounds of / on the corresponding sets K;. We might call the upper bound of the sums (1.1) interpreted in this way the lower Lebesgue integral of the function f over Ky. But actually, this process is of practical importance only for a class of functions, called measurable, and for these the number obtained as the upper bound of the sums (1.1) is called simply the definite Lebesgue integral of f over Ky. What is important, is that the fune- tions which are measurable in the sense of Lebesgue, and whose definition is closely related to that of the measurable sets, form a very general class, ‘This class includes, in particular, all the fune- tions integrable in the Riemann sense. Apart from this, the method of Lebesgue is not only more general, but even, from a certain point of view, simpler than. that of Riemann-Darbonx. For, it dispenses with the simultancous introduction of two extreme integrals, the lower and the upper. Thanks to this, Lebesgue’s method lends itself to an immediate extension to unhounded functions, at any rate to certain classes of the latter, for instance, to all measurable functions of constant sign (cf. below § 10). Finally, the Lebesgue integral renders it per- missible to integrate term by term sequences and series of functions in certain general cases where passages to the limit under the in- c4 CHAPTER 1. ‘The integral in an abstract space. tegral sign were not allowed by the earlior methods of integration. ‘The reason for this is to be found in the complete additivity of Le- besgue measure. The fundamental theorems of Lebesgue (cf. below § 12) stating the preciso circumstances under which term by term integration is permissible, are justly regarded by Ch. J. dela Vallée Poussin [I, p. 44] as one of the finest results of the theory Lebesgue’s theory of measure has, in its turn, led naturally to further important generalizations. Instead of starting with area, or volume, of figures, we may imagine a mass distributed in the Euclidean space under consideration, and associate with each set as its measure, (its “weight” according to Ch. J. de la Vallée Poussin (I, Chap.VJ; 1}), the amount of mass distributed on the set. This, again, leads to a generalization of the integral, parallel to Lebesgue’s, known as the Lebesgue-Stieltjes integral. In order to present # unified account of the latter, we shall consider in this chapter an additive class of measurable sets given a priori in an arbitrary abstract space. We shall suppose further, that in this class, a completely additive measure is determined for the mea- surable sets. These hypotheses determine completely a corresponding method of integration in the Lebesgue sense. All the essential prop- erties of the ordinary Lebesgue integral, exeept at most those im- plying the process of derivation, remain valid for this abstract integral. From this point of view, in a more or less general form, the Lebesgue integral has been studied by a number of authors, among whom we may mention J. Radon [1], P. J. Daniell [2], 0. Nikodym [2] and B. Jessen {1}. For further generalizations (of a somewhat different kind) see also A. Kolmogoroft [1], 8. Boehner [1], G, Fichtenholz and L. Kantorovitch [1], and M. Gowu- rin [1]. §2. Terminology and notation. Given two sots 4 and B, we write AC B when the set A is a subset of the set B, i. e. when every element of A is an element of B, When we have both AC B and BCA, i. e. when the sets A and 8 consist of the same ele- ments, we write A= B. Again, ae 4 means that a is an element of the set A (belongs to A). By the empty set, we mean the set without any element; we denote it by 0. A set A is enumerable if there exists an infinite sequence of distinct elements a, ay, ..., du... consisting of all the elements of the set 4.G21 Terminology and notation. 5 Given a class U of sets, we call sum of the sets belonging to this class, the set of all the objects each of which is an element of at least one set belonging to the class QU. We call product, or common. part, of the sets belonging to the class U, the set of all the objects that belong at the same time to all the sets of this class, We call difference of two sets A and B, and we denote by 4 — B, the set of all the objects that belong to A without belonging to B. Given a sequence of sets (A,) — a finite sequence Ay, Ag .y Any or an infinite sequence Ay, Ay....,4,,... — we denote the sum by 3A, by X'Ay or by Ay + Ag+ 7 nm - A,, in the finite case, and by ZA, by SA, or by 4,4+4,+..+4,+.. in the infinite case. Similarly, merely replacing the sign ¥ by 1, we have the expres- sion for the product of a sequence of sets. If the sequence {4,! is infinite, we call upper limit of this sequence, the set of all the ele- ments @ such that ae A, holds for an infinity of values of the index n. The set of all the elements a belonging to all the sets A, from some n {in general depending on a) onwards, we call lower limit of the so- quence {A,. The upper and lower limits of the sequence {Auli 12.5 we denote by lim sup Ay and lim inf A, respectively. We have (2.1) lim inf 4, = _ HSA . 0 ink tim sup 4,. Jf lim sup A,.=lim inf 4, the sequence {4,) i8 said to be convergent; its upper and lower limits are then called simply limit and denoted by tim Ay. Tf, for a sequence {A,} of sets, we have 4,7 4,1, for each n, the sequence {A,) is said to be ascending, or non-decreasing; if, for each n, we have AuisC Aw the sequence {4,} is said to be descending or non-inereasing. Ascending and descending sequences are called monotone. We see directly that every monotone sequence is con- vergent, and that we have lim.A,—YA, for every ascending se- quence {A,), and lim A, = 1A), for every descending sequence |A,). Finally, given a class € of sets, we shall often call the sets belonging to €, for short, sets (€). The class of the sets which are the sums of sequences of sets (€) will be denoted by €.. The class of the sets which are the products of sueh sequences will be denoted by €) (see F. Hausdorff [U, p. 83).6 CHAPTER I. ‘The integral in an abstract space. § 3. Abstract space Y. In the rest of this chapter, a set X will be fixed and called space. The elements of -Y will be called points. If A is any set contained in X the set V— A will be called com- plement of A with respect to NX; the expression “with respect to X”? will, however, generally be omitted, since sets outside the space X will not be considered. The complement of a set A will be denoted by CA. We ovidently have, for every pair of sets A and B, (3.1) A—B-= A-CB, and for every sequence {X,} of sots WX,=C30X, LN, =CMOX,, (3.2) = : : : limsup X, =CliminfX,, — liminf X,, = Climsup CX,. In the space Y we shall consider functions of a set, and funetions of a point. The values of these functions will always be real numbers, finite or infinite. A function will be called finite, when it assumes only finite valu ‘To avoid misunderstanding, let us agree that when infinite fune- tions are subjected to the elementary operations of addition, subtrac- tion ete., we make the following conventions: a-+-(+00)( ++00)--a= too for a= F 00; (+ 00)-+(—90) = (—o0) + (+00) = (-b00) — (+00). 4+ ($00)=(t00)- a=-foo and a-(00)=(-bo0) a= Foo, according as a>0 or a<0; 0-(400)=(+400)-0=0; af(--o0)=0; a/0=+-00. We call characteristic function ¢,(#) of a set B, the function (of 2 point) equal to 1 at all points of the set, and to 0 everywhere else. The following theorem is obvious: (3.3) If t E,, and Ej-Ey=0 whenever i4:k, then e,(x)= If {By} is a monotone sequence of sets, the sequence of their characteristic functions is also monotone, non-decreasing or non-in- creasing according as the sequence \E,} ix ascending or descending. If {Ey} is any sequence of sets, A und B denoting its upper and lower limits respectively, we have e,(a)=lim sup cy (2), and — e,(:r) = lim inf ey («)s so that, in order that a sequence of sets \B,) converge to a set E, it is necessary and sufficient that the sequence of their character- istic functions {¢, («)) converge to the function {e,(a)}.4 Additive classes of sets. 7 ‘A function assuming only a finite number of different values on a set K is called a simple function on B. Tf My Pgy.005 Pn ave all the distinct values of » simple function j() on a set B, the funetion (2) may on # be written in the form Ha=S v00,(2) where B= By+Ey+.. +E, and By Ky=0 for i4:j. The function / given by this formula over the set / will be denoted Dy {Oy Byj yy Bos 0165 tay Bae ‘The notion of characteristic function is due to Ch. J. de la Vallée Poussin [1] and [I, p. 7h $4. Additive classes of sets. A class X of sets in the spac will be called additive if (i) the empty set belongs to &, (ii) when a set X belongs to & so does its complement CX, and (iii) the sum of a sequence {1} of sets selected from the class &, belongs also to the class &. The classes of sets, additive according to this definition, are sometimes: termed completely additive. We get the definition of a of sets additive in the weak sense if we replace the condition (iii) of the preceding definition by the following: (iii-bis) the sum of bro sets belonging to & also belongs to The sets of an additive class & will be called sets measurable (X), or, in accordance with the definition given in § 2 (p. 5), simply sets (%). We seo at once that, on account of the conditions (i) and (ii), the space Y, as complement of the empty set, belongs to every additive class of sets. Making use of the relations (2.1), (3.1), and (3.2), we obtain immediately the following: (4.1) Theorem. If & is an additive class of sets, the sum, the product, and the two limits, upper and lower, of every sequence of sels measurable (X), and the difference of two sets measurable (%), are also measurable (%). Tn later chapters we shall consider certain additive classes of sets that present themselves naturally to us, in connection with the theory sure, in metrical or in Euclidean spaces. In the abstract space V, about which we have made practically no hypo- thesis, we can only mention a few trivial examples of additive classes of sets, such as the class of all sets in X, or the class of all finite or enumerable sets and their complements. Let us s one further general theorem:8 CHAPTER 1. The integral in an abstract space. (42) Theorem, Given any class W of sets in X, there evists always a smallest additive class of sets containing M. i. e. an additive class Ny DM contained in every other additive class that contains DN. Vor lot %y be the product of all the additive classes that con- tain M1. Such classes evidently exist, one such class being the class of all sots in X, We see at once that the class 3) thus defined has the required properties. §5. Additive functions of a set. In the rest of this chapter we suppose that a definite additive class & of sets is fiwed in the space X. In accordance with this hypothesis, we may often omit the symbol & in our statements, without causing any ambiguity. A funetion of a set, %(X), will be called additive function of a set (X) on a set H, if (i) B is a set (X), (ii) the funetion D(X) is defined and finite for each set XC Ay measurable (&), and if (iii) O(2 X,) = YO(X,) for every sequence (X,} of sets (X) ‘con- tained in # and such that X,-X,—0 whenevor i -k. For simplicity, we shall speak of an “additive function” instead of an “additive function of a set (%)” whenever there is no mistaking the meaning. An additive function of w set (%) will be called monotone on B if its values for the subsets (X) of E are of constant sign. A non- negative function 0(%) additive aud monotone, will also be termed non-decreasing, on account of the fact that, for each pair of sets A and B measurable (X), the inequality ACB implies 0(B) = = WA) + %B—A)> OA). For the same reason, non-positive monotone functions will be termed non-inereasing. (5.1) Theorem, If Me) is an add (5.2) Plim X,,) =lim (X,,) ¢ function on a set B, then for every monotone sequence {X,\ of sets (%) contained in B. If B(X) i8 a non-negative monotone function, then (5.3) O(lim int X,)u—1. For let us choose E,—# and suppose the sets B, for n=1,2,..., & defined so as to satisfy the conditions (6.2). By the second of these conditions with n= k, there exists a measurable set4(C By sueh that (6.3) |D(A)| 2 |OCEy)| 4 ke A)=co, we have only to choose Eyi1=A in order to satisfy the conditions (6.2) for m=k+J. If, on the other hand, W(; A) is finite, we must have W(®; B,— A) + co, and, by (6.3), |O(B,—A)|>|0(A)|—|®(By| Sk, 80 that the conditions (6.2) will be satisfied for n—k+1, if we choose Beyi=B,—A. The sequence {H,) is thus obtained by induction. Now, on account of Theorem 5.1 and of the third of the con- ditions (6.2), we should have the equality #(lim#,)—limO(E,) = If Wo; and since every additive function of a set is, by definition, finite, this is evidently impossible. Q. B.D. It follows from the theorem just proved that every function G(X) additive on a set H is not only finite for the subsets (%) of E, but also bounded; in fact, the values it assumes are bounded in modulus by the finite number W(; £).Ts 6) ‘The variations of an additive funetion i) Theorem 6.1 can be further completed as follows: (6.4) Theorem. For every function F(X) additive on a set By the variations W(D;X), W(®;X) and W(;X) are also additive functions of a set (%) on E, and we have, for every measurable set XCE (6.5) D(X) = WO; X) + W(O; X). Proof. To fix the ideas, consider the function Q(X) = W(@; X). Since this function is finite by Theorem 6.1, we have to show that for every sequence {Y,) of measurable sets contained in 2, and such that Y;-Y,=0 whenever i+ k, (6.6) O (2 X,) = F(X). For this purpose, let us observe that for every measurable set XC SX, we have @(¥)=3O(¥-N,) <2Q(X,), and henee (6.7) QF Xi) STON). On the other hand, denoting generally by Y, any measurable set variable in X,, we have 2,(2X,)20(3Y,) =2O(Y,), and therefore also Q,(2X,)>22(X,). Combining this with (6.7) we get the equality (6.6). Finally, to establish (6.5), we observe that for every measurable subset Y of X we have @(¥)=0(¥)—0(X— Y)<0(X)—W(9;X), andso W(; X)<0(X)—W(; X).Similarly W(®; X)>0(X)—W(G; X). These two inequalities give together the equality (6.5), and the proof of Theorem 6.4 is complete. It follows from this theorem that every function of a set D(X) additive on a set E is, on EB, the difference of two non-nogative additive functions. ‘The formula (6.5) expresses, in fact, D(X) as the sum of two variations of ®(X), of which the one is non-negative and the other non-positive; this particular decomposition of an additive function of a set will be termed the Jordan decomposition.12 CHAPTER I. ‘The integral in an abstract space. We can now complete Theorem 5.1 as follows: (6.8) Cheorem, If O(X) is additive on a set E, we have (lim X,) =lim O(X,) for every convergent sequence \X,\ of sets (X) contained in E. In fact, making use of the Jordan decomposition, we may restrict ourselves to non-negative functions O(X), and for these ‘Theorem 6.8 follows at ones from the second part of Theorem 5.1. §7. Measurable functions. Given an arbitrary condition, or property, (7) of a point 2, let us denote generally by E{(V)] the set of all the points w of the space considered that fulfill this condition, or have this property. ‘Thus, for instance, if f(@) denotes a function of a point defined on a set B and a is a real number, the symbol (7.1) Elw eB; fle) >a] denotes the set of the points « of H at which fl) >a. A function of & point, fw), dofined on a set H, will be termed measurable (X), or simply function (%), if the set H, and the set (7.1) for each finite a, are measurable (). It is easy to see that! (7.2) In order that a function jl) be measurable on a measurable sel Ey it suffices that the set (7.1) should be so for all values of a be- longing to an arbitrary everywhere dense set R of real numbers (the same holds with the set (7.1) replaced by the set B(weH; f(w)>a)). In fact, for every real a, the set R contains a decreasing sequence of numbers {7,3 converging to a. We therefore have Blw eB; f(w) > a]= Y Blve B;f(w) >7,] and, each term of the sum : is on the right being measurable by hypothesis, the same holds for the sum itself (ef. Theorem 4.1). Kvery function (c) measurable on a sot Z, can be continued in various ways, so as to become @ measurable function on the whole space V. For definiteness, wo shall understand by the ea tension of the function f(x) from the set B to the space X, the fune- tion f(x) equal to f(w) on F and to 0 everywhere else. For brevi we shall often deal only with functions measurable on the whole(71 Measurable functions. 13 space X, but it is easy to see that all the theorems and the reasonings of this, and of the succeeding, § could be taken relative to an arbitrary set (&). The equations Bia) a, Bie BY(@) iG Bila) < +00) SEU), B[f(7) > ce] = N ELfl@) > — nh, Bffla)= +00] CEU) <4 co], EL fle — 99] = CBE) > 20] show that for every measurable function f(z) and for every number 4, the left hand sides are measurable sets. Conversely, in the definition of measurable function, we may replace the set (71) by any one of the sets BEf(a) >a], Blfl)a}= Salerro) CELla) n, the functions j,() thus defined are evidently simple and non-negative, and, on account of Theorem 7.3, measurable on E. Further, as is casily seen, the sequence {f,(cv)} is non-decreasing. Finally lim/,(2) =f(«) faa) = \ = n for every xeH; for, if f(*)<+oo, we have, as soon as n exceeds the value of f(x), the inequalities 0 ga), Bf) > yla)] and Bif(x) = g(a), are measurable, The proof follows at once from the identities gey =F Pp » He) >e BUS I=CK>f and Bif=g]=BY Sy) BSN. (8.2) Theorem. If the junction f(r) tx measurable, |f(w)|" is also @ measurable function. For a>0, the proof is a consequence of the identity ELM)" > a) = BUfla) > a!) + BE fle) ately >*S4] for a>o, Ble fo) + A>a)- Bley <4] for a<0, valid for every function f(w) and for all numbers a, ¢-£0, and 6, show, in the first place, that @. f(e)-+@ is a measurable function, if f(w) is measurable. It follows further, from Theorem 8.1 and from the identities: Blef-pg>a~elt>—F | for a>0 Btefeta>a—Bli<—lg 4S] jor aco that if f(z) and g(w) are measurable functions, so is a - fw) + 6 - g(a). (8.4) Theorem. The product of two measurable functions f(w) and g(x) is a measurable function. Measurability of the produet /-y is derived by applying Theorems 8.2 and 8.3 to the identity fg=4L(f+9)?—(f—g)"} the com- plotion of the proof, by taking into account possible infinities of f and g, being trivial. (8.5) Theorem, ¢ the functions n a sequence of measurable junctions \pu(z)', upper bound f,(2), lower bound f,(«), lim sup f(a) and limint f.(c) are also measurable. The measurability of A(x) = upper bound j,{w) follows from the identity Efh(a) >a] = Y BLf.(w)>a]. For the lower bound, the corresponding proot is derived by change of sign. Hence, the functions —hn(x)=upper bound [fv.1(@), fose(), +] are measurable, and the same is therefore true of the funetion Tim sup fu (a) = lim ha(w) = lower bound hi (a). By changing the sign of fu(@), we prove the same for lim inf,16 CHAPTER 1. The integral in an abstract space. §9. Measure. A function of a set (X) will be called a meas- ure (2), if it is defined and non-negative for every set (), and if H(2 Xu) = LaLa) for every sequence {X,} of sets (¥) no two of which have points in common. The number «(X) is then termed, for every set XY mea- surable (&), the measure (#) of X. If every point of a set H, except at most the points belonging to a subset of E of measure (1) zero, possesses a certain property V, we shall say that the condition V is satisfied almost everywhere (1) in EB, or, that almost every (uw) point of E has the property V. We shall suppose, in the sequel of this Chapter, that, just as the class & was chosen once for all, & measure # corresponding to this class is also kept fixed. Accordingly, we shall often omit the symbol (i) in the expressions “measure (u)”, “almost everywhere (#)", ete. Clearly 1X) in sup «(X,), 80 that, in particular, if the sequence |X»\ converges and its sum has finite measure, lim «(Xy) = «(lim X,). Proof. For an ascending sequence {Xn}.-1s,.. the equation lim «(X»)= «(lim X,) follows at once from the relation vx, mt lim X, X43 (Ani Xn), and if the sequence {X,} is descending and «(X,)--oo, then the measure 4(X) is an additive function on the set X, and consequently the required result follows from Theorem 5.1. In exactly the same way, if for an arbitrary sequence (X,} of measurable sets, \'X,, is of finite measure, the measure «(X) is an additive function on this set, and the two inequalities (9.2) and (9.3) follow from Theorem 5.1, To establish the first of these inequalities without assuming that the sum of the sets X, has finite Measure, we write as in the proof of Theorem 5.1 Since the sequence is ascending, and Y,C X, for every », we have (a (lim int Xn) = w (lion Yn) = Hina y(n) < tim ink w(Xn). We conclude this § with an important theorem due to D. Ego- roff, concerning sequences of measurable functions (cf. D. Ego- roff [1], and also W. Sierpinski [3], F. Riesz [2;3], H. Hahn (1, pp. 556—8]). We shall first prove the following lemma: (9.4) Lemma. If E is a measurable set of finite measure (u) and if {fu(a)) is @ sequence of finite measurable functions on E, con- verging on this set to a finite measurable function f(a), there existe, for each pair of positive numbers £, n, a positive integer N and a mea- surable subset H of E such that u(H) <1 and (9.5) [fo (a) — fla)| for every n>N and every weE—H. 5. Saks, Thoory of the Inteeral :18 CHAPTER 1. ‘The integral in an abstract space. Proof, Let us denote generally by H,, the subset of # con- isting of the points « for which (9.5) holds whenever »>m. Thus defined, the sets H,, are measurable and form a monotone ascending sequence, since for each integer m, we have By ff Blve Bs [fe)— h(a] <4 Further, since {fn(«)} converges to f(«) on the whole of B, we have E=)E,, and so, by Theorem 91, «(B)=limy(Bn), ie limp(Z—E,)=0, and therefore, from a sufficiently large 7m» onwards, «(Z—E,,) 0, a subset Q of BE such that n(H—Q)0 a set HC H such that M(H) <€/2" and an index N,, such that (9.7) |fn(a)—F(@)|< 1/2" for n>Ny and for we L—Hm Let us write Q= Y Hy. We find (BQ) < S u(Ha) <¥ ea" =e, a, m1 mn and sinee the sequence j,(7) converges uniformly to f(w) on the set Q on account of (9.7), the theorem is proved.[g 10) Integral. 19 ‘The theorem of E.goroff can be given another form (ef. N. Lusin (I, p. 20), and, at the same time, the hypothesis concerning finite measure of can bo slightly relaxed. (9.8) Tf Bis the sum of a sequence of measurable xets of finite measure («) and af {fuler)} is a wequence of measurable functions finite almost everywhere on this set, converging almost everywhere on H} to a finite function, then the set B ean be expressed as the sum of a sequence of measurable sets H, Ky Hy... such that w(H)=0 and that the sequence {fa(:0)} converges uniformly on each of the sets Bn. For the proof, it suffices to take the ease in which the sot 12 is itself of fin- ite measure. With this hypothesis, we can, on account of Theorem 9.6, define uence {Hr}i-1,.., of measurable sets such that (1! Bx) Any by induction ase Fat} and that the sequence {/a(z)) converges uniformly on the set B7 for each k. Choosing H=E—~™ Bp, we have (H)-=0, and the theorem is proved. ran As we may observe, the hypothesis that the set 2 is tho sum of a sequence of sets of finite measure, is essoutial for the validity of Theorem 9.8. For this pur- pose, lot us take as a spaco X, the interval (0, 1], and as an additive class of sets, that of all subsets of X,. Further, let us define a measure 1, by writing tif X)= co whenever the set Ye, is infinite and jg(X)= x, if X is a finite set and n denotes the number of its elements. The sets of measure (175) zero then i the empty set. Finally, let {gn(a)} be an arbitrary sequence of fun- on this interval, coincide ctions, continuous on the interval [0, 1], converging everywhet but not uniformly ou any subinterval of (0, 1]. To justify our remark concerning ‘Theorem 9.8, it suffices to show that the interval X= [0,1] is not representable as the sum of a sequence {fn} of sets such that the sequence of functions {9 1(e)} s uniformly on each of them. But if such a decomposition were to exist, we might suppose firstly — since the funetions gn (x) are continous — all the sets Hn closed. Then, however, by the theorem of Baire (ef. Chap. IL, § 9) one of them at least would contain a subinterval of [0, I]. This gives a contradiction, since by hypothesis, the se. quence {gn(r}) does not converge uniformly on any interval whatsoever. on § 10. Integral. If we are given in the space X an additive class of sets & and a measure 4 defined for the sots of this class, we can attach to them a process of integration for functions of a point. In fact: (i) If f(@) is @ function (X) non-negative on a set Z, we shall understand by the definite integral (%, 1") of f(@) over E the up- per bound of the sums : pmeelBes where {Hp},-12,...n iS an arbitrary finite sequence of sets (X) such that B=E,+B,+..48, and K;-E,=0 for i--k, and where Yq for k=1,2,....%, denotes the lower bound of f(r) on Ey. oe20 CHAPTER I. The integral in an abstract space. (ii) If f(a) is an arbitrary function measurable (&) on a set B, we shall say that f(w) possesses a definite integral (X, u) over B, if one at least of the non-negative functions f(w) and —f(2) (cf. § 3) possesses a finite integral over H according to definition (i). And, if this condition is satisfied, we shall understand by the definite integral (%, 1) of the function j(w) over # the difference between the integral of jw) and that of —f(«) over B. The definite integral (&, #) of f(@) over E will be written (&) / f(a) du(a). Tf this integral B is finite, the function j(@) is said to be integrable (%, u). For every function f(@) possessing a definite integral over a set H, we evi- dently have (x) [ fay = (%) [fau —(%) | (Hp du = (&) | faw+(% ff dn. We see at once that the two definitions (i) and (ii) are eo m- patible, i. 6. that they give the same value of the integral to any non-negative measurable function. Moreover: (10.1) Lf g = (yy X45 ay X05 ++} my Xmj is a simple non-negative func- tion on the set B=X,+ X,+..+ Xm, the sets X; being measur- able (%), then [ gan = Soin(Xy. 3 es For, if (19... 18 an arbitrary subdivision of H into a finite number of sets (X) without points in common, and if %; denotes the lower bound of g(a) on Ej, we have w; 0, for, otherwise, we should have x [hdp, and, since jf maum [gude, we could replace £8, ne, the set E by the set B—H, on which h(a) does not vanish any- where. Further we shall assume first that 0, <_-+ 09. Let us choose an arbitrary positive number ¢<.7,, and let us denote, for each positive integer x, by Q, the set of the points x of B for which g,(x)>h(z)—«. The sets Q, evidently form an ascending sequence converging to H, and, by Theorem 9.1, we have (Qu) ->4(B). This being so we have two cases to distinguish: (i) «(B) +F cc. We thon can find an integer mp such that for n> we have (HB —Q,)< & and therefore, by Lemma 11.1, [oane [ gud [ [h)—e]dn (x) = Qn Qn =| hae — Qn f hd —ay-tt( EQ) —€ (Qu) => | hdp—[0m + aE) 5 bn B and, passing to the limit, making first m—>o-, and then «> 0, we obtain the inequality (11.5). (ii) w(B)=00. Then, since [ gud é lim [g,du—=co, so that: the inequality (11.5) is evidently satisfied. “iy r—#) (Qn), We obtain Suppose now v= -+ ee. Then by (10.1) and by what has already been proved, Jim i dnd > oH(En)+ ora for any finite number v, and consequently for v= +0o= 0», also; whence, in virtue of (10.1) the inequality (11.5) follows at: onee. (11.6) Lemma. If the junctions of a non-decreasing sequence {yulx)} are simple, non-negative, and measurable (&) on a set B, and if g(a) = lim g,(x), then lim | gn(w) du (a) = | gw) de (a). a vg i Proof. Let Ej, E,,..,Em be an arbitrary subdivision of 2 into a finite number of measurable sets, and let 0,02)... 0m be the lower bounds of g(a) on these sets respectively. Let us write P= (04, Ly3 ay Hy} oj Pan Hui. We evidently have Tim ga (we) =g(a)> 00) on #, and hence, by Lemma 11.4 and by ‘Theorem 10.1[gy Fundamental properties of the integral. 23 lim [god > | yd = So,n(B). 7] ‘ aA It follows that lim a gud > ] gau, and since the opposite inequality is obvious, the proof is ‘complete. We are now in a position to generalize Lemma 11.1 as follows: (1.7) Theorem. Lhe relation (11.2) holds for every pair of functions, gla) and W(x), non-negative and measurable () on the set H, and the relation (11.3) holds jor every function f(a) non-negative and measur- able (%) on the set A+B, where A and B are sets (X) without points in common. Proof. By Theorem 7.4 there exist two non-decreasing so- quences (g,(w)} and {h,(a)) of simple non-negative functions mea- surable (X) on #, such that g(x) = tim gy(x) and Ar) = Timm hala). Now, by Lemma 11.1 (19), we have [ (ga hu)die faite + [hada and hence, making n-»oo, we obtain, on account “of Lemma 11. 6, the relation (11.2). Similarly, if we approximate to /(e) on A+B by a non-decreasing sequence of simple non-negative functions and make use of Lemma 11.1 (2%, we obtain the relation (11.3). (11.8) Theorem, 1° Por any function measurable (%), the integral over @ set of measure zero is equal to zero. 2° If the functions gl) and Max) measurable on a set E are almost everywhere equal on B, and if one of the two is integrable on B, so és the other, and their integrals over B have the same value. 3° If a function f(a) measurable (%) on a set E has an integral over B different from +00, the set of the points w of E at which fla) = +400 has measure zero. In particular, if the integral of lw) over B is finite, the function f(x) is finite almost everywhere on BE. Proof. We obtain at once part 1° of this theorem by making successive use of the definitions (i) and (ii) of § 10. As regards 2°, it is evidently sufficient to consider the case of non-negative functions g(a) and he). If we denote by BE, the set of the points # of # at: which g(a) h(w), we have by hypothesis #(B,) = 0, and, on account of (1°) and of Theorem 11.7, we obtain J ga = | gdu= [ hdu = | hdn, as required. 2k, a é24 CHAPTER I. The integral in an abstract space. Finally, as regards 3°, let us suppose that for a function /(«) measurable (%) on E we have f() = +00 onaset EyC E of positive measure, We then have [ fd > [7 du2n-n(B,) for every n, and a 4 so | fdw=-+co. Consequently, the integral of f(w) over B, if it i exists, is positively infinite, and this completes the proof. We now generalize Lemma 11.1 (1°) and also complete The- orem 11.7, as follows: (11.9) Theorem of distributivity of the integral. Every linear combination with constant coefficients, a-g(w) + b-h(x) of two func- tions g(w) and h(x), integrable (X, ") over a set B, is also integrable over B, and we have (11.10) [(ag + bhydn =a | gdu+b [hau é # é Proof. By Theorem 11.8 (3%), the set of the points at which either of the functions g(«) and h(2) is infinite, has measure zero, and if we replace on this set the values of both functions by 0, we shall not affect the values of the integrals appearing in the relation (11.10). We may therefore suppose that the given funetions g and h are finite on E, Further, the relations [ agdu =a gan, [ohau =o [ han k B e being obvious, we need only prove the formula (11.10) for the caso a=b=1. Finally, the set # can be decomposed! into four sets on each of which the two functions g(a) and h(x) are of cons- tant sign. So that, on account of Theorem 11.7, we may assume that the functions g(x) and h(w) are of constant sign on the whole set E. Now, by the same theorem, the relation (11.11) la th)du = | gdu +if hdu ki i é holds whenever the functions g and h are both non-negative or both non-positive on #, and it only remains, therefore, to show that this relation is valid when g and h have, on E, opposite signs, the one, g(x) say, being non-negative, the other, h(x), non-positive.i] Fundamental properties of the integral. 25 This being so, lot Z, and B, be the sets consisting of the points « of B for which we have g(x)+4-h(x)>0 and g(2)-+h(z) <0, respectively. The functions g, g--h, and —h are non-negative on E, and we therefore have, by Theorem 11.7, [gdu= [g+h)dut [(—h)du=f (g+h)du— [hdu. i iy B, é, i, Similarly —fhau = (=A) du = [(—g—h) du +f gdu =—| (g-+h) du + [ gdu. Therefore, for i=1, 2, we have /(g+h) du =) gdu +-[hdp, and & ot by ‘Theorem 11.7 we obtain the relation (11.11). (11.12) Theorem on absolute integrability, 1° In order that a function f(«) measurable (%) on a set E should be integrable (%, 1) on BH, it is necessary and sufficient that its absolute value should be so. 2° If, for a function g(x) measurable (%) on a set E, there exists a function h(w), integrable (X, 1) and such that |g(a)| and 1«(B) = 0, Neither the theorem thus established, nor its proof, contains, at bottom, anything new, as compared with the similar result for the classical processes of integration of Cauchy, or of Riemann. We now pass on to the proof of theorems more closely related to Lebesgue integration, Among these theorems, a fundamental part is played by the following one, which is due to Lebesgue: (12.3) Theorem. Let f(a) = frlw) be a series of non-negative functions Wat measurable (X) on a set HE. Then (12.4) [ Ha)du(x) =S [ fale) dele). ® aly Proof. From Theorem 11.7, we derive in the first place, that [fans [8 trdul=S [frau tor every m, and so (12.5) | fdu>® | fd. : my To establish the opposite inequality, let us attach, in accordance with Theorem 7.4, to each function j,() a non-decreasing sequence {g'?(a)\e-19,.. of simple functions measurable and non-negative on E, in such a manner that lim g!(«)=f,(e) for n=1, 2,... Let us write mg 12, * sx(a) =, ga). ‘The funetions (2) are clearly simple, measurable, and non-negative, on B, and they form a non-decreasing sequence. Further, for each m, and for k 2m, we have Y (2) < 81 (0) < f(a). Making ko, we derive S f(x) E,, then (12.8) [fdu =X [fan ke "by for every function f(w) possessing a definite integral (finite or infinite) over B. Proof. It is clearly sufficient to prove (12.8) in the case of a function f(#) non-negative on E. Supposing this to be the case, lot us write f,(w)=/(w) for weH,, and f,(e)=0 for eeL—E,. We then have f(«)='f,(#) on H, and, the functions 7, being measurable and non-negative, we may apply Lobosgue’s ‘Theorem 12.3. ‘This gives, by Theorem 11.7, ffde= Sf frdu=Z [trav = 2 fax. QED. i oi ay If a function f(*) has a definite integral & 4) over a set B, then /(«) also has a definite integral over any subset of H mea- surable (%). We may therefore associate with it the function of a set (%) defined as follows: (12.9) F(X)= | flw)dule) where XCE and Xe&. i[p12] Integration of sequences of functions. 29 The latter will be called the indefinite integral (%, «) of f(a) on E. It follows from Theorem 12.7 that, whenever the function f(«) is integrable (2, ) on B, its indefinite integral is an additive func- tion of set (%) on ZB. We end this § with two simple but important theorems. The first is known as Fatow’s lemma, and appears for the first time (in a slightly less general form) in tho classical memoir of P.Fatou [1, p. 375] on trigonometric series. The second is due to Lebesgue [5, in particular p. 376], and is called the theorem on term by term integration of sequences of functions; ef. also Ch. J. de la Vallée Poussin (1, p. 449—453], R. L. Jeffery [1] and T. H. Hilde- brandt [2]. (12.10) Theorem (Katou’s Lemma). If {f.(@)) 4s any sequence of non-negative functions measurable (SX) on a set E, we have | tim int j,,(a) du (w) kt lim inf [ f,(a) du (a). : & Proof. Let us write g,(«) = lower bound [f:(2), fis-1(@), fix2(@), -] where i=1,2,.... Thus defined {g:(m)} is a non-decreasing sequence of non-negative functions measurable on B, and converges on the sot Eto lim inf f,(«). We therefore have, by Lebesgue’s Theorem 12.6, [ima inf f.(«) du(@) = lim [ gi(@) du (w)

[ lim int f, dp, (12.12) " : lim sup [ fadu < [ limsup f,.du. a aap If, further, the sequence if, converges on B to a function f, the sequence is integrable term by term, i. ¢. we have (12.13) lim [ f.du = [ fan. "a é30 CHAPTER [. ‘The integral in an abstract spac Proof. Let g(x) =lim inf f,(e) and let h(«) = lim sup f,(a). We may clearly suppose s(#) <+-0c throughout H. We then derive from Fatou’s Lemma 12.10, lim int { (s+ f.)de> / (s+ g)de and lim int { (s—f,)du > | (s—h)du, whieh giv ca é tions (12.12). Further, if lim j,(w)=f(#), we derive from (12.12) the relation i at once the rela- lim int [ f.du> [ fay > limsup | f,du which gives the equality a r ny (12.13). $18. Absolutely continuous additive functions of a set. ‘The fact that the indefinite integral of a function integrable (%, #) on a set H is, on E, an additive function of a set (X), raises the problem of characterizing directly the additive functions expres- sible as indefinite integrals. If we restrict ourselves to the Lebesgue integral of functions of a real var- iable, we may regard indefinite integrals as functions of an interval, or, what comes to the same thing, as functions of a real variable. In that ease, a neces- sary and sufficient condition for a funetion to be expressible as the indefinite integral of a real function was given, in 1904, still by Lebesgue (I, p. 129, foot- note). A little later (in 1905), G. Vitali [1] explicitly distinguished the class of functions possessing the Lebesgue property by introducing the name of “ab- solutely continuous functions”. The condition of Lebesgue and Vitali was later extended io funetions of a set by J. Radon [1] (ef. also P. J. Daniell (2). But Radon considered only additive functions of sets measurable in the Borel sense in Euclidean spaces, and only measures determined by additive functions of intervals (ef. below Chap- ter III). The final form of the condition of Lebesgue-Vitali, as given in ‘Theorem [4.11 below, is due to O. Nikodym [2] An additive function of a set (%) on a set B, will be said to be absolutely continuous (%,#) on BH, if the function vanishes for every subset (%) of H whose measure (u) is zero. An additive function P(X) of a set (X) on a set E [will be termed singular (%, 1) on E, if there exists a subset By B measurable (%), of measure (1) zro, such that ®(X) vanishes identically on H—E,, i. e. O(X) = @(By-X) for every subset X of H measurable (). The following statements are at once obvious:{§ 13] Absolutely continuous additive functions of a set. 31 (13.1) Theorem. 1° In order that an additive function of a set (%) on a set E should be absolutely continuous (X, #) [or should be singular] it is necessary and sufficient that its two variations, the wpper and the lower, should both be so. 2° Kvery linear combination, with con- stant coefficients, of too additive functions absolutely continuous [or singular] on a set B is itself absolutely continuous [or singular] on EB. 8° Tf a sequence {®,(X)} of additive functions, absolutely continuous [singular] on a set E, converges to an additive function ®(X) for each measurable subset X of B, then the function O(X) is also absolutely continuous [singular]. 4° If a function of a set () is additive and absolutely continuous [singular] on a se E, the function is so on every measurable subset of BE. 5° If BE where \E,} is a sequence of sets (%), and if an additive function MX) on E is absolutely continuous [singular] on cach of the sets Hy, the function is absolutely continuous [singular] on the whole set E. 6° An additive function of a set cannot be both absolutely conti- nuous and singular on a set H, without vanishing identically on B. For sots of finite measure, it is sometimes convenient to apply the following test for absolute continui (13.2) Theorem. In order that a junction OX) additive on a set E of finite measure, be absolutely continuous (X, #) on B, it is necessary and sufficient that to each ©>0 there correspond an n>0, such that u(X)t, Where 1 is a fixed positive number. Let us write H,=lim sup H,. For every m, we then have (By) <3 (En) < 1/2" ', and therefore 1(Hy)=0. On the other in hand, by Theorem 5.1, we have ®(H,) > lim sup ©(H,) > %. This is a contradiction, since O(.{) is absolutely continuous, and the proof is complete.32 CHAPTER I, The integral in an abstract space. (13.3) Theorem. In order that a function ©(X), additive on a set B, be singular (&,u) on B, it is necessary and sufficient that for cach &>0 there ewist a set XCH measurable (%) and fulfilling the two conditions w(X)-0 for every measurable set XC P and G(X)<0 for every measurable set XC B—P. Proof. Vor each positive integer n, we choose a set B, such that 0(H,) > W(0;2)—1/2". By Theorem 6.4 we then have, (14.2) W(9;B,)>—1/2" and WO; B— Bn) S12". Writing P=liming £,, we see that H—P=lim sup(B— Bn) C Su 1) for every m, and therefore, by (14.2), W(0; E—P)< Y WO; B— Bn) < aon oH a which gives W(®;#— P)=0. On the other hand, the lower var- iation W(®; X) is a non-positive monotone function of a measurable set XY CH, and, by Theorem 5.1 and the first inequality (14.2), wo must have the relation |W(; P)| 0, a decomposition of E into a sequence of measurable sets without common. points, Hy, By By, ny... uch that w(H) = 0 and that (14.4) a. (m1) W(X) 0 for every measurable set XC Ay and O(X)—an-"(X)<0 for every measurable set Y(CH—An. Write Bn= Y' 4p. Any measurable Xm where en X, are measurable sets, XC dz for kan, n-41,..., and X;-Xj=0 for + j; and so ®(X)=S O(X,) > S'ak-u(Xs)Sann(X). We obtain 1H 1S subset Y of Bn may be represented in the form X thus a descending sequence of measurable sets {B,} such that OX) >San-w(X) if XCBiy, Xe&, (14.5) O(X) OBEY) 9°" (RA) wh Be. Be), and 2 ea EO. BE) > OE. HY) 9-" « (m—1)- BO”), from which it follows that (2n—k-+1) «(Bs”- By"*)>0, and that (k—2n-+-2) (EO EE?) > 0, Henee p(s”. Bf"')=0 whenever either k>2n+1, or k<2n—2. We may therefore write (14.9) BOC BSED BEAD + BG BS? + Qe? where w(Qn”)=0. This being so, let H— Sams § y Qn". We write f™(e)=2-™.(n—1) for veEy"—H, n=1, Dyan and 0) = for weH. We thus obtain, a sequence {f(«)} of non-negative functions measurable (X) on the set E, By (14.9) we have clearly |/@"*(«)—/™a)|<2-™ on E, so that the sequence {f"(«)} converges uniformly on E to a non-neg- ative measurable function (a). The set IZ being of measure zero, we have, by (14.7) and (14.8), for every measurable set XCE and for every positive integer m, O(X)> aaa (NA) W(X BO) = G(X H) + f fordu, and ] OX)< OX. H)+x 2 mone X- BO) = OX ID) + f fordu +2-™- W(X).[g14] ‘The Lebesgue decomposition of an additive functis 35 Henee, making m-roo, we derive ®(X)= [ f(X)du(w)+0(X-H). ‘ This decomposition, so far established subject to the hypo- thesis that the set is of finite measure, extends at once to sets expressible as the sum of an enumerable infinity of sets of finite Y'A,, where the A, are sets (X) without i common points and of finite measure, then, by what we have al- ready proved, there exists on An a non-negative function f,(«) in- tegrable on A,, and a measurable set H,(CA, having measure zero, such that O(Y.A,)= /fndu+O(X-H,), for n=1, 2,... Lf we #4, now write H=Y'H,, and f(x)=f,(x) for we A,, we obtain a measur- measure. In fact, if F able set HCE of measure zoro, and a function f(X), non-negative and integrable on H, such that, by Theorem 12.7, for every meas- urable set CE (14.10) @X)=N f fdu+ NV OUX- Hy) = [ fdu+ O(X-H). XAn x Now, the indefinite integral vanishes for every set of measure zero, and therefore is an absolutely continuous function; on the other hand, we have %(X-H)=0 for every measurable seb XCE—H. Thus, since the set H has measure zero, formula (14.10) provides a decomposition of @(X) into an absolutely continuous function and a singular function, Finally, to establish the unicity of such a decomposition, suppose that O(X)=9, (X)+0,(X)=V(X)+-4,(X) on B, the functions YX) and 6,.X) being absolutely continuous, and the functions %,(X) and 6,(X) being singular. Then U(X)—V,(X)= =0,(X)—9(X) identically on E, whence by Theorem 13.1 (2° and 6°), we have U(X)=X) and 6,(X)=4,(X), and this com- pletes the proof of our theorem. ‘The expression of an additive function as the sum of an ab- solutely continuous function and of a singular function will be termed the Lebesgue decomposition. The singular function that appears in it is often called the function of the singularities of the given function, From Theorem 14.6, we derive at once *36 CHAPT ER I. The integral an abstract space. (14.11) Theorem of Radon-Nikodym. If E is a set of finite measure, or, more generally the sum of a sequence of sets of finite mea- sure (Wt), then, in order that an additive function of a set (X) on E be absolutely continuous on B, it is necessary and sufficient that this function of a set be the indefinite integral of some integrable function of a point on E. ‘The hypothesis that the set H is the sum of an ab most enumerable infin. ity of sets of finite measnre, plays an essential part in the assertion of the theorem of Radon-Nikodym, just as in Theorem: 9.8. To see this, let us take again the interval [0, 1] as our space Xj, and let the class &, of all subsets of [0, 1] that are measurable in the Lebesgue sense (cf. below Chap. III) be our fixed additive class of sets in the space X;. A measure 1, will be defined by taking jry(X)=co for infinite sets and y,(X)=n for finite sets with n elements. This being so, the sets (%,) of measure (st,) zero coincide with the empty set, and therefore, every additive function of a set (X,) on X, is absolutely continuous (&,, 4). In particular, denoting by A(X) the Lebesgue measure for every set X e,, we see that A(X) is absolutely continuous (%,,'4,) on X,. We shall show that A(X) is not an finite integral (%,. /) on X,. Suppose indeed, if possible, that A(X) = f gle) dete) i for every set eS, the function gle) being integrable (%,, 44) on X,. Since A(X) is non-negative, we may suppose that g(r) is so too, Let H=E[g(s) > 0) and Ep=Ely(z)> 1m] for m=, 2... We have AY, — #)= [ gdu,=0, so that * ra able. MB 1 the sane must be true of En, for some positive integer ng. ‘Thus 1 and this requires the set # to be non-enun ince B= Buy A(En) = | glee) dp (0) = ty (Ba) ftg =" 5 Fug which is evidently a contradiction. § 15. Change of measure. Any non-negative additive fune- tion »(X) of a set (%) may clearly be regarded as a measure cor- responding to the given additive class X. When such a function »(X) is defined only on a set H, we can always continue it (ef. §5, p.9) on to the whole space. The terms measure (»), integral (X,”) ete. are then completely determined for all sets (%), but, in this case, it is most natural to consider only the subsets of F for which the function »(X) was originally given.[§ 15] Change of measure. 37 (15.1) Theorem, Whenever, on a set E measurable (&), we have (15.2) (XX) = (%) | gle)au(w) + 9X) t where 9(X) is a non-negative junction, additive and singular (X,w) on EB, and where glx) is a non-negative function integrable (%, 1) over B, then also (15.3) (%) f fw) dow) = (%) f He) g(a) du(w) + (&) f Hx) de) x i i for every set XCE measurable (&) and for every function f(a) that possesses a definite integral (¥,v) over X. If, further, the function f(a) is integrable (¥,v) over B, the formula (15.3) ewpresses the Lebesgue decomposition of the indefinite integral [ fdr on B, corresponding to the measure u, z the function 6(X)= [ fd9 being the function of singularities (%,) é of the indefinite integral | fav. x Proof. We may clearly assume that f(«) is defined and non- negative on the whole of the set H. We see at once that, for each set ¥CE measurable (%), [ ¢,(a)dv(a) =» ¥)= | ga)du(w) + 9( ¥) = i ¥ | eye) glx) du(w) + | ey(x) d9(x), and hence also that for every i finite function h(x) simple and measurable (X) on a set XC Fy (15.4) [ Wac)dv(a) = | h(x) g(a) dua) + f h(x) aa). i i x Let now {h,(@)} be @ non-decreasing sequence of finite simple functions measurable (¥) and non-negative on X, converging to the given function f(x). Substituting h,(2) for h(a) in (15.4) and making n->oo, we obtain (15.3), on account of Lebesgue’s Theorem 12.6. If, further, /(x) is integrable (%,) over Z, the identity just es- tablished shows at once that the product f(«) g(x) is integrable (%, #4) over H and hence, that the indefinite integral |/gdu is absolutely continuous (%,#) ‘on E. On the other hand, the function 6(X) vanishes on every set on which the function 9(X) vanishes, and therefore, is singular (2,2) on H together with 9X). This completes the proof.38 CHAPTER I. The integral in an abstract space. The wide scope of Theorem 15.1 is due to the fact that, if 1(X) and »(X) are any two measures associated with the same class & of measurable sets and we have at the same time u(L) <+ co, and »(B) <0, for a set Be, then the measure v can be repro- sented on E in the form (15.2), where g() is a function integrable (&,") over the set Hand 9(X) is a non-negative function, additive and singular (%,“) on the same set (cf. Th. 14.6). Hence, with the above hypotheses and notation, in order that f fd» = / fg du x x should hold identically on E, it is necessary and sufficient that the indefinite integral / {dv be absolutely continuous (&, «) on FE. g ‘This condition is clearly satisfied whenever the measure +(X) is itself absolutely continuous (%, /).