Matthews v Williams

[1941] HCA 32

65 CLR 639

MATTHEWS

WILLIAMS AND OTHERS

RESPONDENTS.

PLAINTIFFS AND DEFENDANTS,

ON APPEAL FROM THE SUPREME COURT OF WillConstruction Issue - Whether restricted to childrenDescendants of two

By the will of M., who died in 1867, the period of distribution of his residuary estate (called his " 'trust estate' ") was fixed as at the death of his last surviving child. The general plan of disposal of the trust estate provided, until the death of the last-surviving child, for the division of the income amongst M.' S children, substituting upon the death of each up to the last survivor, the children of the deceased's child as the recipients of the share of income which their parent otherwise would have received. The will then declared that when and so soon as the whole of his children should be dead the trustees should call in investments and sell the trust estate and should hold the trust estate " "upon trust for such of the issue then living of any child or children of mine who being a son or sons have attained or who shall attain the age of twenty-one years or being a daughter or daughters have attained or shall attain that age or have been or shall be married as tenants in common in a course of distribution according to the stocks and not to the number of individual objects the issue of deceased children taking by substitution as tenants in common the respective shares only which their deceased parent would if living have taken and should there be no child or children issue of any or either of my children living at the period of distribution of my trust estate then I give devise and bequeath the same unto my nephews and nieces sons and daughters of my two sisters" (naming them) " in equal shares and proportions as tenants in common and not as joint tenants." The testator was survived by seven children the last of whom died on 5th April 1940. Each of the seven surviving children had children (grandchildren of the testator) who

65 CLR 640

survived the testator. Before the period of distribution some of these grand- children had died. Of those who had died, some had left children (great grandchildren of the testator). Of these great grandchildren some died before the period of distribution leaving children (great great grandchildren of the testator) who were living at the period of distribution. Two grandchildren of the testator (cousins of each other) had intermarried. Both were dead at the period of distribution but three of their children (great grandchildren of the testator) were then living. All seven children of the testator had issue living at the period of distribution.

(1) That as it did not appear that the word issue was intended to be used in the secondary or restricted meaning of "children," the trust should be construed as a stirpital limitation to lineal descendants of the testator, unrestricted in point of degree of propinquity and

(2) That issue who were descendants of two stocks were entitled to share by way of representation or substitution in both the interests which their mother and their father would have taken.

Rule in Sibley v. Perry, (1802) 7 Ves. 522 [32 E.R. 211], discussed. Decision of the Supreme Court of South Australia (Richards J.) In re Matthews, (1941) S.A.S.R. 9, reversed.

APPEAL from the Supreme Court of South Australia.

Thomas Matthews, who died in September 1867, by his last will dated 17th October 1865, after appointing executors and trustees, inter alia, provided as follows:-He gave devised and bequeathed unto his trustees all his real and personal estate upon trust to pay to his son James John Hurd Matthews the rents and profits of a certain farm "during the term of his natural life and from and immediately after his decease I direct that my said trustees shall hold the said farm To the use of such child or children of the said James John Hurd Matthews living at his decease and such issue then living of the child or children of the said James John Hurd Matthews then deceased as either before or after the death of the said James John Hurd Matthews shall attain the age of twenty-one years or dying under that age leave issue living at his her or their death or respective deaths as tenants in common and the heirs and assigns of such child or children and issue respec- tively but SO that such issue shall take only the share or shares which the deceased parent or parents would if living have taken And if no child or issue of the said James John Hurd Matthews living at his decease shall attain the said age or dying under that age leave issue living at his or her death To the use of all and every other of my children share and share alike as tenants in common

65 CLR 641

and not as joint tenants their heirs and assigns for ever I declare that for the purposes of enjoyment and transmission under the trusts hereinafter contained all the rest residue and remainder of my real and personal estate of which I shall die seized or possessed shall be designated my trust estate." The testator then provided for the investment of his trust estate and continued And as to the interest to be procured from the money to arise as aforesaid and the securities whereon the same shall be invested which securities are hereinafter designated 'my trust fund' my said trustees shall stand possessed thereof Upon trust to divide the said interest among all my children in equal shares but subject to the trusts following namely As to the yearly income of each child of mine being a son accruing due in his lifetime Upon trust to pay to him SO much of the same yearly income as would not although the same were payable to him be by his act or default or by operation of law SO disposed of as to prevent his personal enjoyment thereof and to apply much thereof (sic) as would if the same were payable to him be disposed of as last aforesaid for the benefit of his wife children or other issue for the time being in existence or some one or more of the persons who would be his next of kin in such proportions at such times and in such manner as my said trustees shall in their discretion think fit And as to the yearly income of each child of mine being a daughter accruing due in her lifetime Upon trust for such daughter during her life and during any and every coverture of my said daughter to pay the same yearly income as and when the same shall become due and not by way of anticipation into her own hands for her separate use independently of her husband and for which yearly income her receipts shall be discharges to my trustees I declare that when and SO soon as any or either of my children shall die leaving lawful issue (such not being the last surviving child of mine) that my said trustees shall hold the respective share or shares of such rents and interest as aforesaid to which such child or children of mine would have been entitled if living In trust for the mainten- ance and education or otherwise in the discretion of my said trustees to be appropriated for the benefit of my respective grandchildren issue of such child or children SO dying as aforesaid as tenants in common." The testator then declared that when and SO soon as the whole of his children should be dead his trustees should call in and convert his trust estate and should hold the proceeds thereof " Upon trust for such of the issue then living of any child or children of mine who being a son or sons have attained or who shall attain the age of twenty-one years or being a daughter or daughters have attained or shall attain that age or have been or shall be married

65 CLR 642

as tenants in common in a course of distribution according to the stocks and not to the number of individual objects the issue of deceased children taking by substitution as tenants in common the respective shares only which their deceased parent would if living have taken And should there be no child or children issue of any or either of my children living at the period of distribution of my trust estate then I give devise and bequeath the same unto my nephews and nieces sons and daughters of my two sisters Julia the wife of John Vigar of Pitney near Langport Somerset England and Joan the wife of Aaron Larcombe of the same place Farmers in equal shares and proportions as tenants in common and not as joint tenants."

The testator was survived by seven children, and the last survivor of these children did not die until 5th April 1940. All the testator's children married and had children. There were thirty-five grand- children of the testator, of whom fifteen were living on 5th April 1940. Of those who had died thirteen had children, great grand- children of the testator, thirty-eight of whom were still living on 5th April 1940. There were also living on 5th April 1940 fourteen great great grandchildren of the testator who were children of great grandchildren who had died. Of one of the testator's children, no children (being grandchildren of the testator) were living at the period of distribution, though there were surviving great grand- children and great great grandchildren of the testator by that stock. Of the thirteen grandchildren of the testator who had died leaving children, two cousins had intermarried leaving children who survived the period of distribution.

The trustees took out an originating summons for the determina- tion of the following questions :-

1. (a) Whether upon the true construction of the will of the

testator the word " issue" (where used in the trust of the corpus of the estate following the death of the last surviving child of the testator) means children or includes remoter issue (1) on the first occasion the word is used in the said trust, (2) on the second occasion the word is used in the said trust. (b) Whether upon the true construction of the will the word

"children" used in the expression " the issue of deceased children" appearing in the said trust means grandchildren of the testator or includes all or any of the following classes of descendants of the testator -(1) children, (2) grandchildren, (3) great grandchildren, (4) great great grandchildren.

65 CLR 643

(bb) At what stage 'the distribution according to the stocks" H. ('.

directed by the said trust is to begin that is to say- 1 whether it is to take effect in relation to the stocks of descent of all the children of the testator, or 2 whether it is to take effect in relation to the stocks of descent of the six children of the testator who left a child or children living at the death of the last surviving child of the testator, or 3 whether it is to take effect in relation to the stocks of descent of the grandchildren of the testator who survived the period of distribution either by themselves or by the stocks. (c) What persons or classes of persons are entitled to share in

the distribution of the corpus of the testator's estate and in what proportion. (cc) Whether the children of Lavinia Mary Matthews and

Ernest William Matthews both of whom were grand- children of the testator (if they are entitled to share in the distribution of the corpus of the estate) are entitled to take one share as issue of Lavinia Mary Matthews and one share as issue of Ernest William Matthews or only one share and if SO as issue of which stock. These questions were answered in the Supreme Court of South Australia by Richards J. as follows :-

1. (a) (1) and (2) "Issue" means children in both places, i.e., sons and daughters of the testator's children.

(b) "Children" means children of the testator. (bb) The stocks are the seven children of the testator. (c) This is, in effect, answered by the above answers, including that there is an intestacy as to one-seventh of the corpus.

(cc) Calls for no answer. From that decision a great grandchild appealed to the High Court.

Mayo K.C. and Piper, for the appellant and for a respondent great grandchild and a respondent great great grandchild. Assuming Richards J. to be correct in the interpretation placed by him on " issue" where such word first appears in the trust in question, the meaning of the word has been limited by the reference to son or sons and to daughter or daughters (Farrant v. Nichols (1) In re Dean Worland v. Dickinson (2) ). The sub-class that takes "by substitution" is not necessarily ascertained at the period of distribution (Martin v. Holgate (3) ). The words

1(1846) 9 Beav. 327 [50 E.R. 370]. 2(1923) 67 S.J. 768. 3(1866) L.R. 1 H.L. 175. 65 CLR 644

A. by substitution" extend the class by filling in additional donees in

the place of those excluded by dying before the necessary date. MATTHEWS

The clause must be treated as having effect and, in particular, the words "taking by substitution" are inept to indicate merely the quantum that the grandchildren of the testator are to take. The con- struction adopted by Richards J. disregards important words (Davis V. Bennett 1; Beal, Cardinal Rules of Legal Interpretation, 3rd ed. (1924), pp. 614-616). The testator's children took no interest in corpus (Lucas v. Hawkes 2 The "deceased parent" and "deceased children" mentioned in the substitutional clause are persons who would if living have taken as members of the primary sub-class in the trust relating to residue. That is, they must be issue of the testator's children. The testator has used children" and " issue as interchangeable in the relevant trust (Sidle v. Queensland Trustees Ltd. 3 ). The reference in the substitutional clause to 'deceased parent " brings in the rule in Sibley v. Perry 4: See also Hawkins on Wills, 3rd ed. (1925), pp. 114-116. Accordingly, the substituted class will be the children of deceased children of the testator's children. This contention is confirmed by the words in a course of distribution according to stocks and not to the number of individual objects." The presumption is that the stocks are stocks of grand- children (Re Dering Neall v. Beale 5 ). Alternatively, it is submitted that Richards J. was mistaken in holding that issue " where it first appears in the trust of residue means "children" only. The true view is that " issue" means "descendants." The state of the testator's family at the time of the execution of the will must be looked at (Parkes v. Parkes 6 ). The "dictionary" is to be looked for in the will itself (In re Birks; Kenyon v. Birks 7 Edyvean v. Archer; In re Brooke 8 ). The trusts of the farm in which James John Hurd Matthews had a life interest show that " issue" means "descendants." The true scope of the rule in Sibley v. Perry (4) and Pruen v. Osborne 9 is shown in Ross V. Ross 10 and In re Burnham; Carrick v. Carrick 11. "Issue" normally means "descendants" (Davenport v. Hanbury 12; Edyvean V. Archer 13; Ross v. Ross 14 In re Burnham 15 ). There is authority for "children" including " remoter issue" " (Wyth V.

1(1862) 4 De G.F. &J. 327, at p. 2(1920) 28 C.L.R. 266. 3(1915) 20 C.L.R. 557, at p. 560. 4(1802) 7 Ves. 522 [32 E.R. 211]. 5(1911) 105 L.T. 404. 6(1936) 3 All E.R. 653, at p. 662. 7(1900) 1 Ch. 417, at p. 420. 8(1903) A.C. 379, at pp. 384, 385. 328 [45 E.R. 1209, at p. 1210]. 9(1840) 11 Sim. 132 [59 E.R. 824]. 10(1855) 20 Beav. 645 [52 E.R. 753]. 11(1918) 2 Ch. 196, at pp. 201-204. 12(1796) 3 Ves. 257 [30 E.R. 999]. 13(1903) A.C. 379. 14(1855) 20 Beav. 645 [52 E.R. 753]. 15(1918) 2 Ch. 196. 65 CLR 645

Blackman 1 Sidle v. Queensland Trustees Ltd. 2 Knight v. H. C. OF Knight 3 )-compare Farrant v. Nichols 4. The expressions in a course of distribution" and "according to the stocks' indicate a distribution among more than one generation (In re Wilson; Parker v. Winder 5 ). The interpretation placed on the will by the judgment appealed from attributes a most capricious intention to the testator which should, if possible, be avoided: See per Lord Cranworth in Abbott v. Middleton 6. As the children never had an interest in the capital, if only the grandchildren took they would take in succession to, not in substitution for, the children, and the only possible purpose of the "substitutional" clause would be to create an intestacy.

Ross, for certain respondent grandchildren. If "issue" in the trust of residue means something wider than "children," then in answer to question bb it is submitted that the stocks are the grandchildren of the testator who were alive on 5th April 1940, or who had died before then leaving children living on that date. For the prima-facie rule, see Halsbury's Laws of England, 2nd ed., vol. 34, p. 359. In the present will the issue, not the children, are the original takers of the residuary estate. If the residuary gift had stopped at the words as tenants in common," all persons who could bring themselves within the description of "issue" would have taken per capita. The additional words show that they are intended to take not per capita but according to families and that there is to be no competition between parents and children (Robinson V. Shepherd 7; Gibson v. Fisher 8 In re Wilson 9 In re Dering Neall v. Beale 10; In re Alexander: Alexander v. Alex- ander 11 ).

Astley, for a respondent great grandchild. The gift to issue of deceased's grandchildren is an original gift (Lanphier v. Buck 12 ). The class to take is composed of a number of units, each unit being living grandchildren and families of deceased grandchildren. The family of deceased grandchildren can form one unit only (Pruen V. Osborne 13 ). Therefore the descendants who trace their descent through two stocks can take portion of the share of either parent, but not of both (In re Dering; Neall v. Beale (10) ).

1(1749) 1 Ves. Sen. 197 [27 E.R. 2(1915) 20 C.L.R. 557. 3(1884) 10 V.L.R. (Eq.) 195. 4(1846) 9 Beav. 327 [50 E.R. 370]. 5(1883) 24 Ch. D. 664, at p. 667. 6(1858) 7 H.L.C. 68, at p. 89 [11 7(1863) 4 De G.J. &S. 129 [46 E.R. 8(1867) L.R. 5 Eq. 51. 9(1883) 24 Ch. D., at p. 667 10(1911) 105 L.T. 404. 11(1919) 1 Ch. 371. 12(1865) 2 Dr. &Sm. 484, at pp. E.R. 28, at pp. 36, 37]. 494-496 [62 E.R. 704, at p. 708]. 13(1840) 11 Sim. 132 [59 E.R. 824]. 65 CLR 646

H. N. Tucker, for representatives of nephews and nieces entitled under the gift over. The judgment appealed from is correct, except that the gift over has become operative. The testator contemplated that if any one child should die leaving no children at the date of distribution there should be a gift over. The important words are "should there be no child or children issue of any or either of my children living at the period of distribution." If the testator meant what Richards J. found him to mean, the words "any or either of " are superfluous: Compare Farrant v. Nichols 1. Cases such as Powell v. Howells 2 and Holmes v. Meynel 3 are distin- guishable because of the application of principles and implications of cross-remainders (Theobald on Wills, 9th ed. (1939), p. 622). The possibility contemplated by the testator has become effective. The words the same" in the gift over to nephews and nieces relate back to shares of deceased parents. Alternatively, the gift over is operative as to one seventh share, that is, the share in respect of which there were no children or grandchildren living at the period of distribution. It becomes a contest between testacy and intestacy as regards this share. The whole will must be harmonized, and this is best done by giving to the gift over the meaning claimed and applying the presumption against intestacy.

Ligertwood K.C. (with him Culshaw), for respondent grand- children and representatives of next-of-kin. On the authority of Powell v. Howells (2) and Holmes v. Meynel (3) it is submitted that there are cross-remainders, and the estate should be divided into sixths and not into sevenths. Alternatively the division should be into seven parts, six going to the six families where there were grandchildren of the testator living at the date of distri- bution, and the seventh part to the next-of-kin. Throughout the will "issue" means "children" and " issue of children " means "grandchildren." The description in the trust of residue of " being a son or sons or a daughter or daughters" can only mean a grandchild or grandchildren of the testator (Farrant V. Nichols (1) In re Dean; Worland v. Dickinson 4 ). The persons to take are thus fixed as being grandchildren, and the remaining words of the gift merely describe the nature of the tenancy. The "substitutional" gift merely means that each grandchild takes the share of which his parent enjoyed the income during the latter's life. It is true that the parent would not have taken the corpus of the share, but, if living, such parent would have taken the share

1(1846) 9 Beav. 327 [50 E.R. 370]. 2(1868) L.R. 3 Q.B. 654. 3(1681) Sir T. Raym. 452 [83 E.R. 4(1923) 67 S.J. 768. 65 CLR 647

during his life. This reference to "parent" shows that "issue" is used in the narrow sense of "children" (Sibley v. Perry 1; Pruen V. Osborne 2; Martin v. Holgate 3; Hawkins on Wills, 2nd ed. (1912), pp. 117, 118; Jarman on Wills, 7th ed. (1930), vol. 3, p. 1568). The gift over is on default of grandchildren, which restricts the original gift to grandchildren (Ralph v. Carrick 4 ). A detailed examination of the other parts of the will shows that the word "issue" is used throughout in the restrictive sense of "children." The draftsman seems to regard issue as being the appropriate word when speaking of children of the next generation. On the wording of the gift of the farm which James John Hurd Matthews had a life interest in, see Pope v. Pope 5. On the clause dealing with income between the death of one child and the death of the last surviving child of the testator, see In re Thomas Matthews, Deceased 6 Lucas v. Hawkes 7. The word "issue" being thus restricted in other parts of the will, should be SO restricted in the clause relating to residue (Ridgeway v. Munkitterick 8; In re Birks 9 Underhill and Strahan on Wills, 2nd ed. (1906), p. 287; Jarman on Wills, 7th ed. (1930), vol. 3, p. 1575). If this argument be correct, there is one family where there were no grandchildren at the date of distribution, and there is no disposition of this share. As to this it is submitted that there is an intestacy. The nephews and nieces do not take this undisposed-of share, because they become entitled to nothing except upon the total failure of grandchildren. The gift over to them is of the whole estate, not a share of it. Though the condition of the gift over is ambiguous, the meaning suggested is that most calculated to give effect to the intention of the testator. The fact that the construction results in an intestacy does not show it to be wrong (In re Edwards; Jones v. Jones 10; In re Campbell's Trusts 11 ). In the further alternative, if the estate is divisible per stirpes amongst all the issue of whatever degree, then the division should be on the basis of seven stocks of children, and not on the basis of twenty- eight stocks, corresponding with the total number of grandchildren who were either living at the death of the last surviving child of the testator, or, if dead, left issue then living (Jarman on Wills, 7th ed. (1930), vol. 3, p. 1679 Brett v. Horton 12; Willes v. Douglas 13 Waldron v. Boulter 14; In re Hutchinson's Trusts 15 Inre Camp-

1(1802) 7 Ves. 522 [32 E.R. 211]. 2(1840) 11 Sim. 132 [59 E.R. 824]. 3(1866) L.R. 1 H.L., at p. 184. 4(1879) 11 Ch. D. 873, at p. 883. 5(1851) 14 Beav. 591 [51 E.R. 411]. 6(1919) S.A.L.R. 245. 7(1920) 28 C.L.R. 266. 8(1841) 1 Dr. &War. 84. 9(1900) 1 Ch., at p. 418. 10(1906) 1 Ch. 570. 11(1886) 33 Ch. D. 98. 12(1841) 4 Beav. 239 [49 E.R. 331]. 13(1847) 10 Beav. 47 [50 E.R. 499]. 14(1856) 22 Beav. 284 [52 E.R. 15(1882) 21 Ch. D. 811. 65 CLR 648

bell's Trusts 1; In re Rawlinson; Hill v. Withal 2; In re Erring- ton Gibbs v. Nassam 3 ). The present is a fortiori to the above cases, because here the division is expressly said to be according to the stocks and not per capita. The cases show that the stocks are the tenants for life. This estate is already divided into seven parts in respect of income, and the will says that this division is to be kept as regards capital. Robinson v. Shepherd 4, In re Wilson 5, In re Dering 6, In re Alexander 7 are distinguishable, because in those cases there were original gifts to issue and not life interests to parents. The reasoning in Gibson v. Fisher 8 is preferable to that in Robinson V. Shepherd (4).

von Doussa, for the respondent trustees. Piper, in reply.

Cur. adv. vult.

THE COURT delivered the following written judgment :- In this appeal we are called upon to interpret and apply an ambiguously expressed trust in a will which came into effect in 1867. The testator, who died in that year, left an estate which at the time of his death was valued at not more than £5,000. The period of distribution was fixed by the will at the death of the last surviving child of the testator, an event which did not occur until 5th April 1940. In the intervening seventy-three years the value of the estate grew, until it is now estimated at £82,500. The trust which must be interpreted is that defining the class among whom, and the manner in which, the trust estate is to be distributed. The testator left him surviving seven children, whose ages went from fourteen to thirty-five years. His general plan of disposing of his residuary real and personal estate, called his "trust estate," provided, until the death of the last surviving child, for the division of the income amongst his children, substituting, upon the death of each up to the last survivor, the children of the deceased child as the recipients of the share of income which their parent otherwise would have received. The will then goes on to declare that when and SO soon as the whole of his children should be dead the trustees should call in investments and sell the trust estate. Then comes the trust in question, which is expressed as follows: :-

' to hold the same upon trust for such of the issue then living of any child or children of mine who being a son or sons have

1(1886) 33 Ch. D. 98. 2(1909) 2 Ch. 36. 3(1927) 1 Ch. 421. 4(1863) 4 De G.J. &S. 129 [46 5(1883) 24 Ch. D. 664. 6(1911) 105 L.T. 404. 7(1919) 1 Ch. 371. 8(1867) L.R. 5 Eq. 51. 65 CLR 649

attained or who shall attain the age of twenty-one years or being a daughter or daughters have attained or shall attain that age or have been or shall be married as tenants in common in a course of distribution according to the stocks and not to the number of individual objects the issue of deceased children taking by substitution as tenants in common the respective shares only which their deceased parent would if living have taken And should there be no child or children issue of any or either of my children living at the period of distribution of my trust estate then I give devise and bequeath the same unto my nephews and nieces sons and daughters of my two sisters" (naming them) in equal shares and proportions as tenants in common and not as joint tenants." All seven children of the testator had children living after his death. But before the period of distribution some of their children, the testator's grandchildren, had died. Of those who died some left children, that is, great grandchildren of the testator. Of these some died before the period of distribution leaving children in their turn great great grandchildren of the testator. These great great grandchildren were living when the period of distribution came.

Richards J., from whose decision the appeal is brought, held that the objects of the gift are the grandchildren of the testator, and that remoter issue are not included. He further held that the stocks according to which the distribution is to be made are constituted by the seven children of the testator, and that the trust estate is to be divided into seven equal parts, one of which is to be distributed equally among the children living on 5th April 1940 of each of the testator's children. It appeared that in the case of one of the testator's children, the eldest child, all his children died before that date. Richards J. held that there was an intestacy as to that one-seventh share. The intestacy appears to have been brought about by a construction which his Honour placed on the words "the issue of deceased children taking by substitution respective shares only which their deceased parent would if living have taken."

These words were treated as limiting the share which the children of any of the testator's children could take to one-seventh. Accord- ingly, although the gift is to a class consisting " of issue then living" and is a gift of the whole trust estate, it was considered that as there was "issue," in the sense of children, then living of only six of the testator's children, one-seventh of the trust estate was undisposed of. Why the words in question should be treated as limiting the interest to be taken by each family of grandchildren to one-seventh

65 CLR 650

we have not found it easy to understand. Doubtless the expression, 'the respective shares only which their deceased parent would if living have taken," defines the share to be taken by each family of grandchildren (or remoter issue). It defines it by reference to an hypothesis, namely, that the deceased parent remained alive. On that hypothesis you are to find what share he would have taken. Then his offspring or issue get that and no more. But the construc- tion producing the intestacy seems to have added another hypothesis, namely, that all six of that parent's brothers and sisters also lived, or at all events left issue who survived to the period of distribution.

Only on that added hypothesis would the trust estate be divided in seven parts or shares. We can see no justification at all for importing this further hypothesis or condition into the definition of the share to be taken by each family of children or remoter progeny.

It appears to us that all the words mean is that, in the case of each substitution in turn, you are to take the events as they actually existed, with one exception, namely, that you are to suppose that the particular parent of the children to be substituted was still alive and then for the purpose of ascertaining their share inquire what, in the events as they have actually turned out in all other respects, would have been the interest that parent would have taken.

In our opinion, whether the word " issue" means "children" or includes remoter descendants, the limitation to " issue then living" is an ordinary class gift. It is a gift of the whole trust estate to the persons who qualify as members of the class, and there can be no lapse, that is to say, unless there is no-one at all who qualifies as a member of the class and unless the gift over also fails.

We are unable, therefore, to see how the fact that the word "issue" is construed as meaning "children" can result in an intes- tacy. If "issue" means "children," then in the events that happened the consequence would be a division of the estate into six equal parts, a part for every one of the six families of grandchildren surviving the period of distribution for equal distribution among the members of that family then living.

But we have come to the conclusion that the gift to "issue" is not limited to children of the testator's children, but includes remoter descendants. "Issue" is a word with a clear prima-facie legal meaning. It means descendants or progeny. No doubt " issue" is a flexible word and its prima-facie application may be restricted by any sufficient indications appearing in the documents: Cf. In re Birks 1. But ' the essence of the word 'issue,' which primarily means all descendants, is totality rather than succession'

1(1900) 1 Ch., at pp. 418, 420. 65 CLR 651

(In re Cust; Glasgow v. Campbell 1, per Cussen J., distinguishing

issue" from "heirs of the body")

'The best rule of construction is that which takes the words to comprehend a subject that falls within their usual sense, unless there is something like declaration plain to the contrary" (Church V. Mundy 2, per Lord Eldon, applied to "issue" by Knight-Bruce V.C. in Head v. Randall 3, and by Isaacs J. in Campbell v. Glasgow 4 ).

The word "issue" therefore ought not to receive a secondary or restricted meaning unless, upon a consideration of the whole docu- ment, it satisfactorily appears that it was SO intended.

While the will does contain some indications which, by themselves, might support an inference that " issue " is used for "children," none of these indications is irreconcilable with the natural or wider meaning and, on the other side, one part of the trust is consistent only with that meaning, or at all events with a meaning which lets in issue no less remote than grandchildren of the testator's children. That part is comprised in the words, 'the issue of deceased children taking by substitution as tenants in common the respective shares only which their deceased parent would if living have taken." This passage clearly implies that a "deceased child" would, but for his death before the period of distribution, have taken a share under the trust, and on that footing it directs that a substitution shall be made and the issue of a parent shall take his share. Now it is undeniable that no-one nearer to the testator than his grandchildren could take under the trust. For them no-one closer can be sub- stituted than his great grandchildren, that is, children of his deceased grandchildren. It follows, therefore, that great grandchildren of the testator are comprehended in the gift. But they are not brought in under a distinct limitation amounting to a grant or gift to them. They are brought in by force of part of the description of the course of distribution" among the "issue" to whom the trust estate is limited. They are not substituted as a new class outside the main class and introduced to take by representation in lieu of a prospective member of that class who failed to attain membership. They are dealt with as part of the only class to which the gift is made, namely "issue," and are mentioned in directing the mode of division among the class. Indeed, "a course of distribution' is an expression which almost imports successive generations as donees.

We are unable to regard this part of the trust as a mistakenly expressed reference to the shares which the testator's children took

1(1919) V.L.R 221, at pp. 254, 255. 2(1808) 15 Ves. 396, at p. 406 [33 235 [63 E.R. 101, at p. 103]. E.R. 804, at p. 808]. 3(1843) 2 Y. &C.C.C. 231, at p. 4(1919) 27 C.L.R. 31, at p. 51. 65 CLR 652

in income. It appears to us to be the one part of the trust which is incapable of more than one application. It is true that the word "issue" in the expression " issue of deceased children" might be read down to "children," and that the word "parent" might be relied upon for the purpose. But that doubt or uncertainty of meaning is beside the point that the direction in which it occurs is J. unambiguous in its irreconcilability with an interpretation of the

trust which would confine to grandchildren of the testator the class to whom, under the name of "issue" of his children, the trust estate is limited.

The consideration which appears to us to have most weight amongst those which may be said to tend in favour of restricting " issue" to "children" in the description of the donees is the form of the gift over. It must be conceded that according to its more natural meaning it would amount to a gift over of the whole trust estate if at the period of distribution there were no grandchildren of the testator living.

If the gift over were incapable of any other meaning the trust would then present two opposing indications of the meaning of the word " issue"; one, that portion providing for the representation of deceased grandchildren, logically and verbally inconsistent with

restriction of "issue" to "children," and the other, the gift over, forming what must be felt to be an incongruity with any wider meaning. Perhaps, even in such an opposition, it would be right to allow the prima-facie general meaning of the word "issue" to prevail. But the gift over is not incapable of another meaning. Upon examination it appears to be not unambiguously expressed. It is, we think, susceptible of a construction consistent with the wider meaning of "issue." The words "child or children issue of any or either of my children" are not the same as " no child of any of my children," and they do not necessarily have the same meaning.

The words "child or children" may be used not as importing immediate parentage, and the word " issue" may be inserted to show that a remoter degree of descent than child is included. As will appear, the words "sons" and "daughters" are used in an earlier part of the trust in a sense which does not in our opinion import the parentage of any particular generation of issue, and throughout the trust words of lineal relationship are used, as we think, with what has been called a sliding application, that is, SO that they apply between any two successive generations toties quoties. In such a context we think that the gift over should be understood as speaking of the failure of any final class of children

65 CLR 653

descended from the testator, and not as referring to the failure of H. C. his grandchildren.

It is, however, convenient, before dealing with the further con- siderations which on one side or the other affect the interpretation to be placed upon the trust, to state in full the meaning which, in the result, we place upon it. We construe it as a stirpital limitation to lineal descendants of the testator, unrestricted in point of degree of propinquity, but limited to those living at the period of distribu- tion who attain full age or, being females, marry. We think that the relative clause "who being a son or sons" &. has for its antecedent the words "such issue" the "such" and the "who" correspond. The "who" might correspond with the "any" in "any child or children of mine," but though formally it might be better to connect them, yet in substance the probabilities of sense are against that construction.

The words " being a son or sons" and "being a daughter or daughters" we do not read as referring to a filial relation with the testator's children to the exclusion of remoter issue of the testator. The reference is to individuals, members of the class denominated "issue," and they are spoken of as sons and daughters because the whole gift is based upon the idea of representation in successive generations by children of deceased parents who would have taken if living.

We interpret the "course of distribution" as referring to the successive substitution of children for a parent whose death excluded him from the class of donees.

It is well settled that a course of distribution according to the stocks implies that descendants of a lower degree are not let in to compete with any ancestor of theirs, parent, grandparent, et cetera, but only by way of representation of a deceased ancestor.

"According to the stocks" means a stirpital distribution going through the stocks, beginning with the testator's children as the stirpes according to which the first division into shares is made. This must be SO unless " issue" means "children."

The words " issue of deceased children" we regard as an expres- sion of general application, and not confined to any particular genera- tions. It is possible in this phrase to limit issue to "children," and in one sense that must really be its meaning. For the object of this part of the trust is to substitute children for parent toties quoties. But we do not think that " issue" should be read as "children" in such a sense as to restrict the application of the provision to children of the testator's children or to children of his grandchildren. Its application is what has been described as

65 CLR 654

"sliding." In the same way "deceased parent " refers to every

given case of the decease of an earlier possible taker leaving children surviving the period of distribution.

The gift over we regard as taking effect only on a total failure of issue at the period of distribution. The subject of the gift over is described by the relative "same." We think the antecedent of this word is "my trust estate," not "respective shares."

In support of the construction by which the word " issue" was restricted to "children," SO that only grandchildren of the testator would be objects of the trust, the main reliance was placed upon four matters: 1 the gift over; 2 the words "their deceased parent " 3 the words " being a son or sons" and " being a daughter or daughters; ; 4 the manner in which the word "issue" is used elsewhere in the will.

We shall deal with these matters in order. (1) To what we have already said about the gift over it is unneces- sary to add, except perhaps to say that though the strength of the argument it provides in favour of a restricted meaning of issue is great it depends upon the prima-facie or natural effect which the words of the gift over produce upon the mind when it is read indepen- dently and without careful study of the gift which precedes it. That effect weakens in face of a close scrutiny and consideration of the preceding trust, and the countervailing argument provided by the passage beginning "in a course of distribution" is seen to be much stronger, because it shows unequivocally that issue at least extends to great grandchildren of the testator. At the same time closer consideration of the gift over shows that its language is not incapable of a meaning reconciling the gift over with the full connota- tion of "issue," the meaning we have already discussed.

(2) The occurrence of the expression "their deceased parent in relation to " issue" led counsel to invoke the rule in Sibley V. Perry (1).

This rule is briefly stated by Mr. Hawkins, as follows :- Where the 'parent' of 'issue' is spoken of, the word issue' is prima facie restricted to children of the parent " (Hawkins on Wills, 3rd ed. (1925), p. 114). Whether there is such a rule, whether it is a sound or an unsound rule, and whether it should be considered as arising from Sibley v. Perry (1) or Pruen v. Osborne (2) are matters on which judicial expressions of opinion have varied: Cf. Ralph v. Carrick (3) and In re Timson; Smiles v. Timson (4).

1(1802) 7 Ves. 522 [32 E.R. 211]. 2(1840) 11 Sim. 132 [59 E.R. 824]. 3(1873) 11 Ch. D. 873. 4(1916) 1 Ch. 293; (1916) 2 Ch. 65 CLR 655

Perhaps the best discussion of the rule will be found in the judg- ment of Skerrett C.J. in Guardian, Trust, and Executors Co. of New Zealand Ltd. v. Ramage 1.

In Whiting v. Moffatt 2 Holroyd J. made a clear statement of the principles, which it is convenient to set out, the more SO because it is illustrated by reference to Ross v. Ross 3. It is as follows :- "The word 'issue' in its sense of progeny is one of ambiguous import. It may be restricted to children, or it may be extended to all the descendants of a common ancestor. According to the ordinary rules for the construction of wills, a gift to the issue of any person is a gift to all the living descendants of that person (Daven- port v. Hanbury 4 ). But if property is given to such of several persons as may be living at the testator's death, or other stated period, with a direction that the issue of any of them who may die before that period shall take the shares which their parents respec- tively would have taken, if living, or words to that effect, or even if the gift be to such of the several persons as may be living at the period fixed, and to the issue of such of them as may be then dead, with a similar direction, in either of these cases the general rule is, that 'issue' must be translated 'children' (Sibley v. Perry 5 Pruen v. Osborne 6; Bradshaw v. Melling 7; Smith v. Horsfall 8 Maynard v. Wright 9 These rules, however, are not rigid, and will yield to slight expressions of the testator's intention, to be collected from other parts of his will. For instance, where the gift to the issue is distinct, and followed by the direction that the issue shall take the parent's share, the question may arise whether the word 'parent' is limited to the specified persons who would, if living, have taken to the exclusion of all issue. Thus in Ross v. Ross (3) there was a bequest to Christian Ross for life, with remainder to all her children who should be living at the time of her decease, and the issue then living of such of her children as might have died in her lifetime each of her surviving children to take an equal share; and the issue, if more than one, of such of her children as might have died in her lifetime, to take equally amongst them the share which their parent would have been entitled to if he or she had survived the said Christian Ross, and if but one, then to take the child's share. There was a gift over on the death of Christian Ross, without leaving a child, or issue of a child. One of Christian Ross's children predeceased her, leaving no children,

1(1927) N.Z.L.R. 288. 2(1898) 20 A.L.T. 113 4 A.L.R. 3(1855) 20 Beav. 645 [52 E.R. 753]. 4(1796) 3 Ves. 257 [30 E.R. 999]. 5(1802) 7 Ves. 522 [32 E.R. 211]. 6(1840) 11 Sim. 132 [59 E.R. 824]. 7(1853) 19 Beav. 417 [52 E.R. 412]. 8(1858) 25 Beav. 628 [53 ER. 776]. 9(1858) 26 Beav. 285 [53 E.R. 908]. 65 CLR 656

but a grandchild, who survived Christian Ross and it was held that the grandchild took a child's share along with Christian Ross's children. The reason of the decision was, that the word 'parent' was not brought into such correlation with the word 'issue' as to indicate that the meaning of 'issue' was restricted, and that there was nothing to limit the generality of the word 'issue' in the gift over" 1.

The foundation of the rule is the relationship implied by the use of the word "parent" between a first or earlier taker or person otherwise referred to and a second or subsequent class of takers called " issue." It is evident that in applying the rule the first step to take is to identify the earlier taker or takers to whom the word "parent" is applied. When this is done and it appears, as of course in most instances it naturally does, that ascertained or ascertainable persons are regarded as standing in the relation of parent and child to the issue, an inference arises that by issue the testator meant children.

But, adopting the language of Cotton L.J., in Ralph v. Carrick 2, if there is nothing to show that the person referred to as a parent must of necessity be the ascertained person the foundation of Lord Eldon's decision upon the third clause of the will in Sibley v. Perry 3 has no application. This point is brought out by Sargant J. in In re Burnham 4. He says: "In Pruen v. Osborne 5, Smith V. Horsfall 6, and other cases of the kind, the only persons-expressly mentioned as donees under the original gift, and for whose decease provision is expressly made are the first generation, therefore natural to limit provisions for representation to the case of this generation, and when this is done, the force of the word 'parent' almost necessarily limits the class of introduced repre- sentatives to the next succeeding generation."

Skerrett C.J. 7 was even more definite. "The so-called rule applies only where the word 'issue' can be clearly referred to the issue of a parent who if living would be the first taker under the gift " and again, " I think that to apply the principle of Sibley v. Perry (3) you must be satisfied that the word 'parent' when used in relation to 'issue' must point to the first taker under the trust."

Now in the present case the donees are a class living at a specified event described as issue, that is, prima facie, successive generations then living. There is to be a stirpital distribution among them. It is quite clear that when a possible taker of an earlier generation,

1(1898) 20 A.L.T., at pp. 113, 114; 2(1879) 11 Ch. D., at p. 886. 3(1802) 7 Ves. 522 [32 E.R. 211]. 4(1918) 2 Ch., at p. 204. 4 A.L.R., at p. 248. 5(1840) 11 Sim. 132 [59 E.R. 824]. 6(1858) 25 Beav. 628 [53 E.R. 776]. 7(1927) N.Z.L.R., at pp. 291, 293 65 CLR 657

because he predeceases the event, does not qualify for membership of the class, the next generation is substituted to represent him, and it is equally clear that the relationship between him and the next generation is that of parent and child. Upon the question whether the earlier generation is confined to a particular degree of propinquity to the testator the use of the word "parent" can throw no light. But it can and does show that there must be at least two generations contemplated by the word " issue," and by doing SO provides a strong reason against the construction adopted by Richards J. confining the class of donees to the grandchildren of the testator.

(3) The expressions "son or sons and "daughter or daughters" no doubt suggest a relationship of parent and child, and as the testator's children are the generation last referred to it is at first a natural association of ideas to treat that generation as the parent. But again, as the clause proceeds it becomes clear that under the term " issue" more than one generation is comprehended. There is therefore less ground than might at first appear for supposing that it is between the testator's children and their children and no other successive generations that the filial relation is contemplated. It is clear that within the class called " issue' the relation of parent and child is contemplated and there is no occasion for thinking that such a relation with a generation antecedent to and outside that class was in the mind of the draftsman.

Farrant v. Nichols 1 was decided on a very different limitation, one in which Lord Langdale found in the expression "whether son or daughters" a restrictive intention. No general rule was laid down, and Lord Langdale observed that "each case depends on the peculiar expressions used, and the structure of the sentences" 2.

The will contains a number of other instances in which the word " issue is used. In one case it is used to cover remoter descendants. In one or two it is doubtful whether it has or has not a restricted meaning. In others the context shows that "children" only were intended.

The fact appears to be that the draftsman employed the word wherever the dominant idea in his mind was lineal descent. Some- times he either by design or by chance qualified the degree of proximity; sometimes he did not.

In our opinion there is no guidance to be obtained from the other instances in the will of the use of the word " issue."

1(1846) 9 Beav. 327 [50 E.R. 370]. 2(1846) 9 Beav., at p. 330 [50 E.R., at p. 371]. 65 CLR 658

The trusts under consideration are like those of the will before Eve J. in In re Sutcliffe Alison v. Alison 1, and it appears from the report 2 that he placed the same construction upon "issue" as we have adopted. Unfortunately his reasons are not reported.

A question has arisen of a peculiar nature in reference to three of the testator's great grandchildren. Their parents, who dead, were cousins, the mother being a daughter of a deceased daughter of the testator and the father a son of a deceased son of the testator. The question is whether they are entitled to share by way of representation or substitution both in the interest which, if she had lived, their mother would have taken, and in that which, if he had lived, their father would have taken. In our opinion the answer must be yes. The purpose of the trust is to distribute the trust estate according to the stocks, not equally throughout. Equality is the standard only as between stocks and as between children of a deceased parent in dividing his share.

Here the three great grandchildren in question have separate titles, SO to speak, to share in the division of different interests. This is the view taken by Sargant J. in In re Burnham 3. The limitation in Pruen v. Osborne 4 which was relied on was quite different and the decision appears to us to have little or no bearing upon this particular question.

In our opinion the appeal should be allowed and the order of the Supreme Court discharged except as to costs and in lieu thereof it should be ordered that the questions in the originating summons should be answered as follows :-

1. (a) The word "issue" includes remoter issue in both cases.

(b) The word "children" in the expression " issue of deceased

children" applies whether the deceased child was a grand- child or remoter issue of the testator. (bb) The distribution is to begin by taking the seven children of

the testator who had issue living at the period of distribu- tion, viz., 5th April 1940. (c) The descendants of the testator living on that date but SO

that children or remoter issue of any descendants of the testator living on that date shall not compete with their parent or ancestor and shall be excluded. (cc) They are entitled to take one share as issue of Lavinia Mary

Matthews and one share as issue of Ernest William Matthews.

1(1934) Ch. 219. 2(1934) Ch., at p. 221. 3(1918) 2 Ch., at p. 206. 4(1840) 11 Sim. 132 [59 E.R. 824] 65 CLR 659

The costs of all parties of and incidental to the appeal should come out of the estate, those of the trustees as between solicitor and client.

Appeal allowed. Order of the Supreme Court discharged

except as to costs, and in lieu thereof order that the ques- tions in the originating summons be answered as follows: 1. (a) The word "issue" includes remoter issue in both cases. (b) The word "children" in the expression "issue of deceased children" applies whether the deceased child was a grandchild or remoter issue of the testator. (bb) The distribution is to begin by taking the seven children of the testator who had issue living at the period of distribution, viz., 5th April 1940. (c) The descend- ants of the testator living on that date but SO that children or remoter issue of any descendants of the testator living on that date shall not compete with their parent or ancestor and shall be excluded. (cc) They are entitled to take one share as issue of Lavinia Mary Matthews and one share as issue of Ernest William Matthews. The costs of all parties of and incidental to the appeal to be paid out of the estate; those of the trustees as between solicitor and client. Solicitor for the appellant, R. N. Finlayson. Solicitors for the respondents, L. von Doussa; Thomson, Buttrose, Ross &Lewis; Genders, Wilson &Pellew; Piper, Bakewell &Piper Baker, McEwin, Ligertwood &Millhouse; J. F. Astley.