Sulaiman and young, JJ.
1. One important question raised in this appeal is that even if the dedication made by Mt. Rajwanti, a Hindu widow, was without the oral authority of her deceased husband, the alienation was only voidable at the option of the next reversioner, and therefore the defendants have no locus standi to challenge the alienation. This is based on the contention that according to the pedigree, as printed on p. 3 of the paper book, defendants 4 and 5 are not the heritable bandhus of the last male owner Suraj Prasad. The latter question is undoubtedly a substantial and important question of Hindu law which requires an authoritative pronouncement. The learned advocate for the plaintiffs strongly relies on the statement of the law as contained in Dr. Sarvadhikari's book on the Principles of the Hindu Law of Inheritance, Edn. 2, pp. 571-2, where he emphasized that if there are two females (on the father's side), these two again must be related as mother and daughter. He also strongly relies on the classification of the cognate sapindas who can be heritable bandhus given by the same learned author on pp. 591-2. Further reliance is placed on the case of Sheo Nandan v. Munni A.I.R. 1923 All. 398 decide by a Bench of this Court. In that case the plaintiff was the great-grandson of the sister of the father of the propositus. The learned Judges relying on the passage in Dr. Sarvadhikari's Principles of the Hindu Law of Inheritance, held that he was not an heir. With great respect we would say that we have some doubts as to the correctness of this decision. It is however pointed out that the test laid down by Dr. Sarvadhikari, and the rule of law in Sheo Nandan's case A.I.R. 1923 All. 398 were approved of by another Bench of this Court in Bam Sia v. Bua A.I.R. 1924 All. 790, where the previous decision of this Court in Shib Sahai v. Saraswati A.I.R. 1915 All. 409 was also explained.
2. On the other hand, there can be no doubt that a Bench of the Madras High Court in Kesar Singh v. Secy. of State A.I.R. 1926 Mad. 881 has expressly dissented from Dr. Sarvadhikari's view. The judgment of their Lordships of the Privy Council in the case of Bamckandra Mart and v. Vinayah Venkatesh A.I.R. 1914 P.C. 1, which was incorrectly reported in 12 A. L. J. 1281, does not reproduce the tests laid down by Dr. Sarvadhikari.
3. As the case is of a high valuation, we think that the following question of law should be decided by a larger Bench:
On the pedigree as given on p. 8 of the paper book, are defendants 4 and 5 heritable bandhus of Suraj Prasad in the absence of any nearer heir.
4. We direct that the record be placed before the Hon'ble the Chief Justice for the constitution of a larger Bench.
5. The question for consideration in this reference is whether the last male owner's father's father's daughter's son's daughter's sons are his heritable bindhus. The admitted 'pedigree is as follows:
Mt. Badama Sheo Ghulam
Snkhdeo Suraj Prasad.
Kosho (Deft. 4) Kali (Deft. 5.)
6. The learned advocate for the claimants contends that inasmuch as Kesho and Kali are within five degrees of the common ancestor Shri Tewari, (when the claimants, their female ancestor and the common ancestor are all included in counting), and Suraj Prasad was within only three degrees of the common ancestor, the case is covered by the rule laid down by their Lordships of the Privy Council in Bamchandra v. Vinayah A.I.R. 1914 P.C. 1. He has to concede that their Lordships laid down the further condition as to the mutuality of sapinda relationship. But he contends that no other restrictions should be placed on heritable bandhus. He relies strongly on the case of Kesar Singh v. Secy. of State A.I.R. 1926 Mad. 881, in which the pedigree was exactly the same as is before us, and the learned Judges of the Madras High Court dissenting from certain conclusions of professor Sarvadhikari held that the claimant was a heritable bandhu. He also adopts the line of reasoning followed by a writer in Vol. 9 of the Madras law Journal p. 54 (notes) onwards. On the other hand the learned advocate for the respondents places his strong reliance on the conclusions arrived at by Professor Saravadhikari in his Tagore Law Lectures (1880) dealing with the principles of the Hindu Law of Inheritance pp. 591-592 and pp. 571-572 (Bdn. 2). His conclusions were:
(1) That the heritable bandhus can belong only to three classes mentioned in the Mitakshara and they fall within five groups namely, (i) the owner's descending line, (ii) owner's father's family, (iii) owner's maternal grand-father's line (iv) father's maternal grand-father's line and (v) mother's maternal grand-father's line; and (2) that there must be at least one female between the given person and the deceased; and that (in the case of bandhus ex parte paterna) there could not be more than two females between them and if there are two females (on the father's side) these two again must be related as mother and daughter to each other.
7. A bandhu is a cognate relation, i.e., one related through a female. The institutes of Manu do not refer to bandhus at all. The claim of bandhus as heirs was apparently first recognized by Yajnavalkya, who lived about the second centuary A.D. The Mitakshara by Vijna-neswara, who flourished about 12th century, is a commentary on Yajnavalkya's text. The Viramitrodaya is a commentary by Mitra Misra on the Mitakshara. These commentaries are of high authority in the Benares School. The relevant texts of the Mitakshara as translated by Colebrooke are as follows:
8. Chapter 2, Section 6 p. 1:
(1) On failure of gentiles, the cognates are heirs. Cognates are of three kinds; related to the person himself, to his father, or to his mother, as is ceclared by the following text. The sons of his own father's sister, the sons of his own mother's sister, and the sons of his own aunt, and the sons of his mother's maternal uncle must be considered as his own cognate kindred. The sons of his father's paternal aunt, the son? of his father's maternal aunt, and the sons of his father's maternal uncle, must be deemed his father's cognate kindred. The sons of his mother's maternal uncles, must be reckoned his mother's cognate kindred. (2) Here, by reason of near affinity the kindred of the deceased himself, are his successors in the first instance; on failure of them, his father's cognate kindred; or if there be none his mother's cognate kindred. This must be understood to be the order of succession here intended.
9. The original words in the texts indicating the three classes are : (1) atma bandhus, i.e., own bandhus (2) pitri bandhus, i.e., bandhus on the father's side or as they are sometimes called the father's bandhus and (3) matri bandhus, i.e., bandhus on the mother's side or as sometimes called the mother's bandhus. It was laid down by their Lordships of the Privy Council in Girdari Lall v. Bengal Government  12 M.I.A. 448 that the enumeration of bandhus in the Mitakshara is illustrative and not exhaustive. It has been further laid down by their Lordships in Muthusami v. Simambedu  19 Mad. 405 and in Aditnarayan v. Mahabir Prasad A.I.R. 1921 P.C. 53 among the three classes, the atma bandhus succeed before pitri bandhus who succead before matri bandhus.
10. Again, it is now also well settled that the bandhu relationship under the Mitakshara is based on propinquity or affinity, i.e., community of blood, or, to be more accurate, particles of the same pinda or body. (Community of blood is not a Hindu conception, as a child is supposed to inherit some materials from his father and some from his mother; and he inherits blood from his mother only: Ghosh's Hindu Law, Vol. 1,190, Edn. 3. In contrast with this, under the Dayabhaga the principle underlying succession is one of spiritual efficacy, based on community in the offering of pinda or funeral cake. It may be noted here that the word 'pinda' means both body and also cake or ball.
11. The leading case on this point is that of Bamchandra v. Vinayek A.I.R. 1914 P.C. 1. It is necessary to give a brief summary of the judgment in this case, as counsel for both the parties rely upon it strongly. Their Lordships quoted the text of Manu that the property of a near sapinda shall be that of a near sapinda' and pointed out that this refers only to the succession of one male to another, for females inherit by express rules. The right of collaterals is dependent on the existence of sapinda relationship between the propositus and the claimant. The author of the Dayabhaga considers that 'sapinda' means
community in the offering of funeral oblations, as pinda literally signifies a ball of rice offered at the performance of obsequial rites.
12. The author of the Mitakshara abandoned the ancient doctrine and construed sapinda relationship to arise from community of blood, i.e., of particles of the same body. The bandhus are distant kinsmen related to the deceased through females. They make their appearance as heirs first in Yajnavalkya's enunciations. Their Lordships quoted from Ch. 3 of the Achara Kanda, the first book of the Mitakshara on established rules of conduct. In laying down that a man should marry a girl who is his nonsapinda it is stated that she is called his sapinda who has particles of body of some ancestor in common with him. Their Lordships then quoted:
Verse 53: 'After the fifth ancestor on the mother's and after the seventh on the father's side;' 'on the mother's side' in the mother's line, after the fifth, 'on the father's side' in the father's line after the seventh (ancestor) the sapinda relationship ceases, et cetera.
I have put the words 'on the mother's side' and 'on the father's side' within quotation marks to indicate that they are mere reproductions from the original text quoted first the comments upon them being in the mother's line and in the father's line respectively. This is made clear in Professor Sarvadhikari's translation of the text, as well as in a Poona translation of the original.
13. Their Lordships had no manner of doubt that in this quotation Vijnaneswara was laying down rules for the limitation of sapinda relationship generally, and not with regard to marriage only. Their Lordships closely examined the doctrine of the Mitakshara, and pointed out that the word 'bandhu' in its system has a distinctive technical meaning signifying the bhinna gotra sapindas. Their Lordships then quoted the relevant texts of the Mitakshara. Referring to the earlier case of Girdhari Lall  12 M.I.A. 448 their Lordships remarked that:
the remark of the Board that the text did not purport to be exhaustive enumeration of all bandhus capable of inheriting and was not cited for that purpose by the author but simply as a proof or illustration of his proposition that there are three kinds or classes of bandhus, hardly warranted the contention that the classes specified by Vijnaneswara can be added to.
14. But in that case it did not seem necessary to their Lordships to enter upon the determination of the question whether the classes can be extended. They held that:
the limitation of five degrees clearly applies and can only apply to the bhinna gotra sapindas, and that the limitation is not restricted to matrimonial affinity only.
15. Their Lordships rejected the opinion expressed by Shastri Golap Chandra Sarkar in his work on Hindu Law that the word 'bandhu' in the Mitakshara means and includes either all cognate relations without any restriction, or at any rate, all cognates within seven degrees on both the father's as well as on the mother's side. Their Lordships concluded that the right of inheritance is founded on sapinda relationship which, under the Mitakshara, means consanguinity; but this bond comes to an end with the fifth degree when the descent is through a female, as the right to inherit could not continue after the relationship on which it is founded has come to an end.
16. The next proposition which their Lordships considered was the doctrine of the mutuality of sapinda relationship, based on the rule enunciated by Manu. They referred to Professor Sarvadhikari's Lectures and to Bhattaoharyya's Hindu Law with approval for the proposition that in order to determine whether any persons are heritable cognates of the propositus it is necessary to see whether they are sapindas to each other. This had been laid down in Umaid Bahadur v. Udaichand  6 Cal. 119. Their Lordships also referred to Babu Lal v. Nanku Ram  22 Cal. 339, where it had been held that a man in order to be a heritable sapinda of the propositus must be so related to him that they are sapindas of each other.
17. The general conclusions to which their Lordships came were in the following terms: (1) That the sapinda relationship on which heritable right of collaterals is founded ceases in the case of bhinna gotra sapinda with the fifth degree from the common ancestor and (2) that in order to entitle a man to succeed to the inheritanceof another he must be so related to the latter that they are sapindas of each other. In the case before their Lordships the claimants were the deceased's paternal grandfather's son's son's daughter's daughter's sons. They were bhinna gotra beyond the fifth degree, and the element of mutuality was wanting. Their claim was accordingly rejected.
18. The learned advocate for the appellant contends that the two tests laid down by their Lordships in Ramchandra's case A.I.R. 1914 P.C. 1 are exhaustive. Their Lordships no doubt laid down that no bhinna gotra sapinda beyond the fifth degree from the common ancestor can be a heritable bandhu on the ground of sapinda relationship. The learned advocate for the appellants urges that it was necessarily held that all persons within five degrees are heritable bandhus. The converse proposition does not follow. There are two circumstances which conclusively show that no exhaustive definition of 'bandhu' was intended to be laid down, and the other restrictions also are possible. First on p. 298 (of 41 I.A.) it was pointed out that females inherit by express rules only. They must therefore be excluded as heirs. Secondly, their Lordships left the question open whether the three classes of bandhus could be added to.
19. It seems convenient now to examine the analysis put forward by Professor Sarvadhikari. It must be noted that the lectures were delivered in 1880, and although in Edn. 2 they have been revised and brought up to date by incorporating the various rulings, the main body of the text could not be substantially altered. There are certain points which compel one to hold that the propositions laid down by Professor Sarvadhikari should not be accepted off hand without a close investigation. At p. 294 (of 41 I.A.) the first couclusion arrived at in Bamohandra v. Vinayek A.I.R. 1914 P.C. 1 is quoted wrongly and it is stated that the sapinda relationship ceases in the case of bhinna gotra sapindas with the fifth degree from the propositus. The report of the judgment as originally received has misprints which were corrected subsequently. The conclusion is obviously quoted from the incorrect copy of the judgment. The editor thought that it fitted in very well with the general scheme put forward by Professor Sarvadhikari. The conclusion, as quoted in the report of the full judgment in the foot-note is however correct. On p. 610 he thought that, if instead of applying the rule of affinity in regulating the order of succession, he had applied the principles of religious benefit, the results would have been exactly the same and that this was an additional proof that the doctrine of religious benefit and the doctrine of affinity are not really different from each other; but they are inseparably connected with each other and when properly applied they yield the same results. This proposition can no longer be said to be quite correct when the order of preference among bandhus has to be determined. Their Lordships of the Privy Council have distinctly laid down in Vedachela v. Subramania A.I.R. 1922 P.c. 33 that in each class propinquity is the governing factor and the nearness in blood determines the right to succeed, though as between the bandhus in the same class and in the same degree of nearness, the spiritual benefit conferredupon the propositus may be a ground of preference. Professor Sarvadhikari's classification of atma bandhus apparently follows the line adopted under the Dayabhaga law, where the capacity for conferring spiritual benefit is the governing test. The pinda is offered to the three immediate paternal ancestors and the three immediate maternal ancestors, the pindalepas are offered to the three paternal ancestors above them. The ascent both through the father and the mother are exclusively in the agnatic lines. He takes the father's agnatic ascendants as well as the mother's agnatic ascendants as the owner's atma bandhus. These he divides into ex parte paterna and ex parte materna, according as they are on the father's or the mother's side.
20. He then takes up the father's own bandhus. He naturally identifies the father's father's agnatic ascendants with the line of the owner's father's agnatic ascendants, and takes the father's mother's agnatic ascendants as the second class, namely, pitri bandhus. Similarly he identifies the mother's father's agnatic ascendants with the line of the owner's mother's agnatic ascendants, and takes the mother's mother's agnatic ascendants as the matri bandhus. The descendants of the ancestors in each class are of course included in that class. In table 1(a) showing the atma bandhus (ex parte paterna) p. 606-A he shows the owner's great-grandfather's daughter's son as an atma bandhu (ex parte paterna). In the same way in table 1(b), p. 607-B he shows the owner's mother's grandfather's daughter's son as-an atma bandhu (ex parte materna). Bub this classification is inconsistent with the Mitakshara and cannot be accepted. In the text of the Mitakshara quoted above the sons of the father's paternal aunt, are described as pitri bandhus, and the sons of his mother's paternal aunt are described as matri bandhus. These are the identical relations given in the two tables.
21. The extension of the atma bandhus in the father's and the mother's lines above the grandparents would make a considerable difference when the order of preference is to be considered, as all atma bandhus come before the pitri bandhus. That the table of succession as given by Professor Sarvadhikari and followed by other-learned authors is wrong as regards the place assigned to the maternal uncle, is clear from the pronouncement of their Lordships of the Privy Council in Vedachela v. Subramania A.I.R. 1922 P.C. 33: vide the last-paragraph but two in the judgment. It follows that according to the Mitakshara the cognates connected through the ascendants above the paternal grandfather just as much as those above the paternal grandmother must be regarded as pitri bandhus. Similarly cognates connected through the ascendants above the maternal grandfather just as much as those above the maternal grandmother must be regarded-as matri bandhus. Atma bandhus are either in the descending line of the propositus or in his line up to the grandparents. As we go upwards from the grandfather in the father's line or the paternal grandmother in her paternal line we get pitri bandhus. If we go above-maternal grandfather in his ancestral line or the maternal grandmother in her paternal line we get matri bandhus. There-are of course certain restrictions which will be mentioned later; bandhus in the father's father's agnatic line are pitri bandhus (ex parte paterna), those in the father's mother's agnatic line are pitri bandhus (ex parte materna), those in the mother's father's agnatic line are matri bandhus (ex parte paterna) and those in the mother's mother's agnatic line are matri bandhus (ex parte materna). These classifications alone appear to be consistent with the text of the Mitakshara. This division is in perfect accordance with that given in the Smriti Ghandrika Ch. 11, Section 5, para. 14 and the Saraswati Vilasa para. 595, which are books of high authority in Southern India, and which are quoted by their Lordships of the Privy Council in Vedaohela v. Subramania A.I.R. 1922 P.C. 33. They show clearly that the descendants of ancestors above the grandparents on the father's and the mother's side are pitri and matri bandhus respectively.
22. It will however be shown presently that in spite of the divergence in the classification, Professor Sarvadhikari's conclusions and the fixing of the five groups are both substantially correct, though his propositions have to be re-stated in new terms in view of the pronouncement by their Lordship's of the Privy Council in Ramachandra's case A.I.R. 1914 P.C. 1.
23. The learned advocate for the defendants-appellants adopting the line of argument taken by the writer in the Madras Law Journal contends that the Mitakshara speaks'of one's own or atma bandhus and one's father's and mother's bandhus, but not one's father's own bandhus or mother's own bandhus. The contention is that the father's bandhus are spoken of generally and should be further divided into father's own bandhus (pitri atma), father's father's bandhus (pitri pitri atma), and father's mother's bandhus (pitri matri atma.) Similarly he would divide mother's bandhus into three sub-divisions, viz., mother's own (matri atma) mother's father's bandhus (matri pitri atma) and mother's mother's bandhus (matri matri atma). In this way the relationship of heritable bandhus would extend beyond the grandparents of the owner and would include all the male and female great-grandparents.
24. It seems to me that the extension is fantastic and is based on an obvious fallacy, that the words 'father' and 'mother' have been used in a generic sense. If this were so then mother would stand not only for mother's mother's mother but also for mother's mother's father, which is untenable. Another error, which would be pointed out later on is the assumption, that the words 'father' and 'mother' have been used in the possessive sense.
25. The language of the text of the Mitakshara itself makes it clear that the bandhus are to be divided into three classes. They are of three kinds only. There is-no justification for adding another class to these three classes. When the text says that they are of three kinds, the three classes must be three, neither more nor less. As has been pointed out by Professor Sarvadhikari on p. 591 even, the additional bandhus mentioned by Mitra Misra fall within the same three' classes.
In fact no commentator within the last eight-hundred years has ever suggested that there may be a class of heritable bandhus other than the three specifically mentioned in the Mitakshara.
26. In Amrit Kumari v. Lakhi Narayan  10 W.R. 78 the Full Bench remarked that Vijnaneswara's enumeration is used simply in establishment of the proposition: that there are three classes of bandhus. The three classes are therefore exhaustive and not illustrative.
27. The case of Kesar Singh v. Secy. of State A.I.R. 1926 Mad. 881 is directly in favour of the appellants. The mistakes in Professor Sarvadhikari describing the father's father's-sister's son and the mother's father's-brother's son and mother's father's, sister's son as atma bandhus contrary to-the text of the Mitakshara where they are called pitri bandhus and matri bandhus respectively, were pointed out and so also the wrong place assigned to the-maternal uncle. The learned Judges thought that for the five groups prepared by Professor Sarvadhikari, there was a support neither in any ancient text or recognized commentary nor on any authority. They considered that there was no-justification for narrowing the class of heritable bandhus however logical and rational the theory may be.
28. If the Madras view were accepted the-necessary result would be to include great-grandparents within the atma bandhus, and father's father's mother's agnatic ascendants as well as mother's-mother's mother's agnatic ascendants may become bandhus. It is urged on behalf of the. appellants that Mr. Mulla has adopted this view in his Hindu Law, p. 48 and. also pp. 52-3 (Bdn. 6). But Mr. (now the Rt. Hon'ble Mr.) Mulla has merely incorporated the result of the decision without expressly saying that it is a right decision.
29. The words in the text of the Mitakshara are pitri bandhus and matri bandhus. They have sometimes been translated as 'father's bandhus' and 'mother's bandhus.' Had the words been 'pituh' bandhu or 'matuh' bandhu, the translations would have been undoubtedly correct. But the words are 'pitri bandhu' and 'matri bandhu,' which, when translated word for word mean 'father bandhu' and 'mother bandhu.' As pointed out by Mukerji.J., these are 'samasa' or shortened forms, and may stand for one out of several forms as the context may demand. That the words 'pitri' and. 'matri' have been used in the adjectival sense and not in the possessive, and they do not mean father's bandhus and mother's bandhus, but the person's bandhus through the father, i.e., on the father's side or in the father's line, and the person's bandhus through the mother, i.e., on the mother's side or in the mother's line, is obvious from the fact that they must still be the person's bandhus and not bandhus of only the father or the mother. The words cannot mean that a cognate who is the father's bandhu or the (mother's bandhu, though not a bandhu 'of the propositus, is an heir. It cannot therefore be urged that the words' father's bandhus' and 'mother's bandhus' include the same atma relations of the father and the mother as the atma relations of the owner.
30. It is well known that the Hindu conception of sapinda relationship is primarily along the agnatic line. When one talks of the father's line, he generally implies the father's father's agnatic line or at most father's mother's agnatic line, and not father's mother's mother's line. Similarly the mother's line would be the mother's father's agnatic ascendants or the mother's mother's agnatic ascendants and not the mother's father's mother's line or the mother's mother's mother's line.
31. As pointed out in Umaid Bahadur v. Udoi Chand  6 Cal. 119, the sapinda relationship has been defined in Achara Kanda. Verse 52 contains the following passage:
So also (is the nephew) a sapinda relation of his maternal aunts and uncles and the rest, because particles of the same body (the paternal grandfather) have entered into (his and theirs); likewise (does he stand in sapinda relationship) with paternal uncles and aunts and the rest.
32. This is followed by Verse 53 which fixes the limit of the fifth ancestor on the mother's side. The commentary on this verse also shows agnatic ascent or descent. The Full Bench in Umaid Bahadur's case  6 Cal. 119 in order to explain the doctrine of the mutuality of sapinda relationship gave an illustration showing that the common ancestor's daughter's daughter's son's son was not a bandhu. Though their statement was not absolutely necessary, it cannot be said that it was altogether an obiter dictum. It is entitled to great weight.
33. In Babu Lal v. Nanhu Ram  22 Cal. 339 (at p. 345) the learned Judges pointed out who according to the explanation or definition given in the Mitakshara was a sapinda. The fourth group was stated in the following words:
Any ascendant (within the fifth degree reckoned from the owner inclusive of himself in the maternal line, that is any of the four maternal ancestors, namely, the mother her father, her grandfather and the rest.
34. Here too the mother's ancestors are in her agnatic line. This is how the learned Judges understood the definition of a sapinda.
35. The owner's bandhu may be in his father's agnatic line or in his mother's agnatic line or he may be in the father's line or the mother's line, which would include the father's mother's agnatic line and mother's mother's agnatic line. They cannot be in the father's mother's mother's agnatic line or in the mother's mother's mother's agnatio line.
36. There is one more circumstance which also conclusively narrows the group of bandhus, i.e., only male ancestors beyond the circle of grandparents. If bandhus were intended to include descendants of all ancestors, males and females, limited by the only restriction that they should be within five degrees, then there would have been no necessity to mention the last two classes, namely, bandhus on the father's side and bandhus on the mother's side. All of them could have been described as bandhus of the owner himself limited by the sole restriction of five degrees. It would then have been quite sufficient to say that all the cognates relations of the owner within five degrees are bandhus. The fact that in addition to the atma bandhus, the two other classes of bandhus are mentioned, shows that the atma bandhus of the writer would not have included his father's mother's father and higher ancestors, nor his mother's mother's father and higher ancestors, although they may be within five degrees of the propositus. Yet another strong circumstance points to the same conclusion. If the second ,and third classes had meant father's bandhus and the mother's bandhus just as the first class, the owner's own bandhug, then as bandhu relationship ceases in the fifth degree, one would have to count to a common ancestor up to the fifth degree from the father and the mother, and therefore up to the sixth degree from the propositus. Ramohandra's case A.I.R. 1914 P.C. 1 has now settled that this cannot be done. There is only one conclusion to which one can come and that is that the father's mother's mother's father and the mother's mother's mothor's father are, not included in one's heritable bandhus. The ancestral lines going outside the grandparents are purely agnatic and not cognatio.
37. It may be noted in this connexion that there is no other authority in support of the contrary view except the opinion of Dr. Golap Chandra Sarkar who thought that the limitation of degrees was applicable to marriage only and not to inheritance and that mutuality of sapindaship was unnecessary, and accordingly concluded that a bandhu may be taken to mean any relation however distant, belonging to a different family whose relationship can be traced: Hindu Law, Edn. 6, 479. This view has been definitely rejected by their Lordships of the Privy Council in Ramchandra v. Vinayak A.I.R. 1914 P.C. 1.
38. It may also be stated that all other modern writers like Mayne Bhattacharyya and Ghose, not to mention living authors like Dr. Gour, have all excluded female ancestors above the circle of grandparents. In an earlier edition of Mr. Mayne's Hindu Law the father's sister's son's daughter's son was included in the table on the authority of Parot Bhupalal v. Hari Lal  19 Bom. 63. But he has been deleted from the Table in Bdn. 9, obviously because the Bombay case is no authority for his inclusion in the table of bandhus. The table excludes father's mother's mother's ancestors and mother's mother's mother's ancestors.
39. We can now group together the atma bandhus of the owner first. These are now well known to be included (1) within the circle of grandparents in the ascending line, and their descendants under certain limitations including certain descendants of the owner himself. For pitri or matri bandhus we go outside the circle and higher, but in accordance with the ordinary conception we must travel upwards along agnatic lines that is (2) the father's father's agnatic ascendants (pitri, ex parte paterna)(3) the father's mother's agnatic ascendants (pitri ex parte materna), (4) mother's father's agnatic ascendants (matri ex parte paterna) and (5) mother's mother's agnatio ascendants (matri ex parte materna), and not in the agnatic ascendants of the mothers of these ancestors.
40. These groups correspond with the five groups of families enumerated by Professor Sarvadhikari, and establish his first proposition. It will however be noticed that the order of precedence is not identical with that given by him. As regards the method of counting the five degrees, Professor Sarvadhikari's statement in Ch. 2, p. 72 on his Hindu Law (Edn. 2) may be conveniently quoted:
Accordingly, it is to be understood that the fifth from the mother is she who is (the fifth) in the line of descent from (any ancestor of the mother, up to the fifth ancestor (and counting her and such ancestor, each as one degree) in the computation beginning with the mother (and counting her as one degree), and the mother's father, paternal grandfather, and the like.
41. As a necessary result of the mutuality of sapindaship, if it involves a similar counting from the side of the claimant also, the seven degrees in the owner's father's line have also to be reduced to five degrees. In view of the ruling in Ramchandra's case A.I.R. 1914 P.C. 1 the bandhus beyond the fifth degree from the common ancestor shown in Professor Sarvadhikari's book or Mayne's book would most; probably not be heritable bandhus. Their Lordships have clearly laid down that the sapinda relationship on which the heritable right of collaterals is founded ceases in the case of the bhinna gotra sapinda with the fifth degree from the common ancestor. On the same principle, the restriction against the inclusion of a great-grandmother will apply in the case of the owner's atma bandhus in the descending line as well.
42. Thus certain propositions are HOW quite clear. (1) As females inherit only under express rules, no female can be a heritable bandhu (2). The common ancestor can always be taken to be a male, because a female anceetor's husband will always be the nearar & an equal ancestor of the descendants of his wife. (3) The relationship in the mother's line ceases with the fifth degree. It follows that an ancestor beyond the fifth degree from the propositus would be outside the circle. (4) the only exception which was recognized before Ramchandra v. Vinayak A.I.R. 1914 P.C. 1, appears to have been the case of the father's mother's great-grandfather, who is included on the theory of what is called a 'frog's leap,' by which the father's mother is excluded in the counting. (5) There must be mutuality of sapindaship, and the deceased and the claimant must he so related that they are sapindas of each other.
43. The mutuality of sapindaship enables us to reverse the process, and applying the same ralou to the claimant, deduce the converse prepositions, viz.: (1) the propositus may be a male or female, for instance, a maiden leaving stridhan property; (2) the common ancestor will always be a male; (3) the claimant must be within five degrees of the common ancestor; (4) the owner's father's mother is counted in calculating the number of the degrees, because she becomes like a daughter in the reverse process and (5) the propositus must be a sapinda of the claimant.
44. From these it is obvious that between the propositus and the claimant there must be at least one female relation; ether-wise the latter would not be a bandhn at all; there however may be more than one female. Indeed, the maximum number of females may be five, viz., the propositus, two females between her and the common ancestor (not treating her as a female), and two more between the common ancestor and the claimant. It follows that between the propositus and the common ancestor on the one side and the common ancestor and the claimant on the other side there can be only throe intermediary relations. We have already Been that, there CMI be only a male ancestor above the grandparents. This is true as regards both sides of the ancestor. This means that there cannot be any mother's mother's mother or mother's father's mother, for in such a case, there would be a female above the grandparents. If these conditions are not full-filled the claimant cannot be an heir.
45. It necessarily follows that the claimant in order to be a heritable bandhu cannot be lower than the common ancestor's daughter's grandson or the common ancestor's son's great grandson the proposition may be stated in a simple way by saying that there should be no great-grandmother in the line on either side of the common ancestor.
46. This conclusion bears out the correctness of the second proposition laid down by Professor Sarvadhikari although it has to be restated in the following form:
There must be at least one female between the given person and the deceased, and there cannot be more than two females between the propositus and the common ancestor on the one side, and not more than two females between the common ancestor and the claimant on the other side, and if two females exist on either side they must be related as mother and daughter to each other. This holds, good in the case of pitri and matri bandhus, whether ex parre paterna or ex parte materna, and also in the case of atma band bus whether in the ascending line or in the descending line.
47. Applying these principles to the case before us it at once becomes obvious that Keshor and Kali cannot be the heritable: bandhus of Suraj Prasad. Their great-grandmother Mt. Badamo intervened. Bandhus cannot come down lower than Sukhdeo's son if any.
48. As I was one of the referring Judges I must point out that our doubt as regards the decision in Sheo Nandan v. Nanni A.I.R. 1923 All. 398 arose because the claimant there was within five degrees of the common ancestor, and the judgment did not give any reasons except relying on the authority of Professor Sarvadhikari's opinion which in its turn proceeded on grounds soms of which as pointed out above, were not sound. In the light of the conclusion now reached, the actual result of the decision in Sheo Nandan's case A.I.R. 1923 All. 398 was correct, and so were the decisions in Ram Sia v. Bua A.I.R. 1924 All. 790 and Hari Har Prasad v. Ram Daur : AIR1925All17 . These cases were in accordance with the ruling in Krishna Ayyangarv. Venkatrama Ayyangar  29 Mad. 115. In neither of these cases any great-grand-mother intervened. My answer to the question referred to us is therefore in the negative.
49. The point of law that has been referred to the Full Bench would require the following pedigree for its proper statement:
Sheo Ghulam Mt. Badamo
Suraj Prasad |
Proposition) Sukhdeo (son)
Kesho Prasad Kali Prasad
(Defendant 4). (Defendant 5).
50. The question for determination is whether defendants 4 and 5 namely, Keaho Prasad and Kali Praia.1, are heritable bandhus to Suraj Prasad and are his heirs in the absence of any nearer heir. This is an abstract question of law, and before I can answer it, I have to see who are the bandhus, according to the Mitakshara law, which governs the case. The Mitakshara Ch. 2, Section 6 lays down as follows:
On failure of gontiles the cognates aro heirs. Cognates are of three kinds, related to the person himself, to his father or to his mother. The word 'cognate' is a translation of the word 'bandhu'.
51. In Ch. 2, Section 5, Clause (3) we find the definition of the term cognate ' or 'bandhu.' It is stated: 'for kiusmon sprung from a different family but oounccted by funeral oblations are indicated by the term 'oognato ' (bandhu)(Colobrooke's Translation).
52. Putting the two definitions together we find that kinsmen sprung from a different family but connected by funeral oblations are to succeed after the failure of 'gentiles.'
53. The expression connected by funeral oblations' used by Mr. Colebrooke is a mistranslation. The word in the original is 'sapinda' which, in the case of Bengal law, may properly be translated by the expression ' connected by funeral oblations,'' but, in the case of the Mitakshara law, the meaning is somewhat different. According to Vijnaneshwara, the word 'pind' means ' body or particles of body.' Without going into the discussions raised and answered by Vijnyanesh-wara, We may take it as settled law that under the Mitkashara law, a sapinda is one who is connected by means of blood.
54. It follows that a bandhu is a blood relation but belonging to a ' different' family.
55. The 'different' family ' is a family which is connected with the propositus through a 'female or females.' A 'cognate' therefore is a blood relation who is connected with the propositus through a female or females. A blood relation as defined above may include almost the whole world and therefore a limitation had to be put to the words 'blood relation' or 'sapinda.' A sapinda relationship has been defined by Vijnaneshwar in that portion of his commentary on the Institutes of Yajnavalkya which deals with Achar or rules of conduct.
56. In Ch. 1, Sloka 53, the sage Yajnavalkya defines who a sapinda is by the' following rule:
Panchamaat saptamaadoordhwam maatritah pitritasthathaa.
57. The translation of the commentary of Vijnanenhwar on this passage, as given by several writers, is not very happy. The writer' of the Mitakshara first puts down the word 'matrito' and then gives its meaning as 'matuh santaane.' The three words ' matrito ' 'matuh' and 'santane' are not to be taken together.
58. As I have said, the first word is taken from the text and the next two words are its meaning. This would indicate that, according to Vijnaneshwar, the word 'matrito' means 'in mother's line'. Then come the words 'panchamaat oordhwam', which moan above the No. 5.' Then comes the word 'pitrito' which is quoted from the text, and it is followed by its meaning given by Vijnaneshwar as 'pituh santaane' which when properly translated would mean in father's line'. Then follow the words 'siptamafc urdhwam', that is to say, above the seventh degree.' Then follow the words 'sapindyam nivartat'. The whole commentary together means:
in the case of a child claiming through its mother and in the case of child claiming through its father the sapinda relationship ends; with the fifth and seventh degree respectively.
59. The translation that we find given by English text-writers is:
On the mother's side in the mother's line after the fifth, and on the father's side in the father's line after the seventh ancestor the sapinda relationship ceases : West and Buhler, Vol. 1 120.
60. This translation however does not affect the true meaning. What Vijnaneshwar means is that where sapinda relationship is to be considered of a person, who claims through his mother we have to count up to the fifth degree, inclusive of the said person, and in. the case of a person claiming through his father, we have to count up to the seventh degree, inclusive of the said person.
61. Vijnaneshwar further lays down that the six paternal ancestors are one's sapindas and so are the six lineal male descendants, himself being counted as the seventh. He further lays down that where the ascending or descending lines branch off, 'santan bhede' the seven degrees are to be counted from where the lines branch off, inclusive of the person from whom the lines branch off. Similarly, it would follow that whore a person is claiming through his mother, and the line ascending or descending branches off, five degrees are to be counted from where the break occurs, inclusive of the person from whom the line branches off. Thus a person who is in the seventh degree, descending, say, from the third male ancestor of the propositus, will be a sapinda (of the propositua). Vijnaneshwar also explains that in the case of One claiming through one's mother, one's sapindas are the mother, the maternal grandfather, his father and so on, all on the male line. The text is as follows:
Tathaacha maataramaarabhaya tatpitripi taamahaadiganangayaam panchamapurushawartinee maatritah panchameetyupacharyate.
62. The definition of sapinda' having occurred in the part of the Mitakshara dealing with rules of achar' or conduct' at one time it was contended that the definition could not hold good for the second part of the book, which dealt With inheritance. But this contention is no longer regarded as valid. Vijaaneshwar himself says : see his commentary on verse 52 of Ch. 1 of the Institutes of Yajnavalkya that the word 'sapinda' is to be understood in the sense explained by him, wheresoever the same may occur. It is a rule of construction of Hindu law texts that a word once used in one sense is to be taken as having been used in the same sense throughout the book, unless a clear intention to the contrary is established. It may be mentioned that their Lordships of the Privy Council, in the case of Ram Chandra v. Vinayak A.I.R. 1914 P.C. 1 have ruled that the word 'eapinda' in the chapter on inheritance has the same meaning as in the chapter on 'achar'.
63. utting together, then, the definitions of all the words with which the term bandhu' has been defined, we find that bandhu is one who is related to the propositus within seven or five degrees, as the case may be, through a female or females. The fact must not be lost sight of that the lines of relationship must be through males and males alone, and in the case of a claim through the mother, it is the mother's father, her paternal grandfather and so on that constitute the line for sapindaship.
64. According to Manu (Ch. 9, V. 187) a sapinda alone can take a sapinda's property. This means that for the purposes of inheritance by a sapinda, both the claimant and the deceased must be related to each other as sapindas. At the outset, this may appear to be a mere truism, but this is not always the case. We have seen that a sapinda relationship, where it is traced through the mother stops at five degrees and when it is traced through the father it stops at seven degrees. It will therefore follow that while A may be a sapinda of B, B may not necessarily be a sapinda of A. I shall give the most familiar illustration to be found in text-books. In the following diagram A is the common ancestor and on the right and left are his two branches of descendants. 8 represents sons and D represents daughter.
65. S 5 on the right hand side being 7th in degree would be S2's (on the left) sapinda, but counting from S5 who claims through his mother D, five degrees would be the limit of his sapindaship and, therefore S5's sapindas would terminate with S2 on the right. In this case S5 cannot inherit to S2, on the left.
66 We have then discovered that there is an additional qualification which a claimant must possess before he can claim, as a bandhu, the property of a deceased person. This qualification is that the claimant and the deceased must be related to each other as sapindas. Now comes a matter which is the most controversial of the whole rule as to inheritance by a Bandhu in Hindu law. Vijnaneshwar says:
Gotrajaabhaawe bandhawo dhanabhaajah. Bandhawascha triwidhaah; aatmabandhawah pitribandhawo matribandhawascheti (2-8-130).
67. This, when translated, means in the absence of a cognate, bandhus get the inheritance. Bandhus are of three kinds, namely, Atmabandhu, Pitribandhu and Matribandhu. Then follows an example, being a quotation.
One's father's sister's song, one's mother's sister's sons and one's maternal uncle's sons are Atma bandhav, one's father's father's sister's sons, one's father's mother's sister's sons and one's father's maternal uncle's sons are pitri bandhav and then mother's father's sister's sons, mother's mother's sister's sons, and mother's maternal uncle's sons are matri bandhavas.
68. I have stated that these are examples given by Vijnaneshwar and are not to ba taken as an exhaustive list of bandhus.' Vijnaneshwar puts this list as if the list were a quotation, which under modern times, would be put within a pair of inverted commas. This follows from the fact that the list is prefaced with the words, 'yathoktam' which mean 'as has been said.'
69. It is too late in the day to contend that the list of the throe kinds of bandhus is exhaustive, for the Privy Council have laid it down in clearest terms that the list is not so : see Girdhari Lal v. Government of Bengal  12 M.I.A. 448.
70. The question however is ; can we add to the three classes of bandhus? The question arose before their Lordships in the Privy Council in the case of Ram Chandra v. Vinayak A.I.R. 1914 P.C. 1 quoted above, but their Lordships did not decide the question although it appears that their Lordships were inclined to the view that the classes could not be added to. My own opinion is that the classes cannot be added to, and, for this reason, that by the statement that bandhus were of three kinds, Vijnaneshwar limited the number of classes. These classes, as already stated, are Atma-bandhus, Pitri-bandhus, and Matri-bandhus. If we could add to the classes there would be no sense in stating that the classes were only three. The fact should not be lost sight of that Pitri-bandhus and Matri-bandhus are the bandhus of the propositus himself and have no status except as the bandhus of the propositus. We have to find out the bandhus of the propositus. We are-told that we would find them under three descriptions, and under three descriptions alone. The text quoted by Vijnaneshwar, to explain the nature of a bandhu doss not at all use the word 'bandhu.' The word that it used is 'bandhav,'' which only means a 'relation' or 'connexion'. The word 'bandhu' has been used by the Mitakshara in a 'technical'' sense, while the word 'bandhus' have been used in a non-technical sense. The commentary of Vijnaneshwar and the text quoted by him, taken together, indicate that 'bandhus' of the propositus are to be found among the persons who were' related directly to the propositus, through a female, and among persons who were similarly related to him through his mother and father.
71. So far as the term 'Atma-bandhu'' goes, it is not capable of having two meanings. It must be taken as meaning ''direct relations'. The expressions, pitri-bandhu' and rnatri-bandhus' ar& compound words, formed by the process-known in Sanskrit grammar, as 'samasa' or 'shortening process.' On breaking up the 'samasa', we may have the worda 'Pitooh bandhu' or 'Matooh bandhu' or we may have the words 'Pitri bandhu' or 'Matri-bandhu'. The two words relating to father and mother having the same form, I will take one only, for the' sake of brevity. Now 'Pitooh bandhu' if we treat the expression as a shashthi tatpurusha samasa' would mean father's, bandhu'. If we treat the expression as. a panchami tatpurusha' the word would mean 'bandhus through or by reason of the fathers', according to the rule 'hetau cha', which means that the suffix: 'panchami' which gives the same forntt as shashthi to the words pitri and matri,, may be used where 'the reason is to be-assigned.' Then again if we treat the words as ' tritia tatpurusha samasa' then by the rule tritia karane' we still have the same meaning, viz., 'through the father'. So it cannot be asserted with any show of reason that the words 'pitri bandhu' and 'matri bandhu' must mean 'father's bandhu' and 'mother's bandhu.' The analysis or breaking up of the words which would give them the mornings, 'by reason of' or 'through' father or mother appeal to me, as the correct rendition. If we adopt the analysis which would mean 'father's or mother's bandhu' we shall find that the classification becomes useless and meaningless, if we should further say that the word bandhu' is used there in the technical sense, and not as having the same meaning as the word 'bandhav' in the text quoted. Then all relations of the propositus, howsoever remote would come in as his ''bandhu' either through himself or through his father or mother or through the paternal and maternal ancestors of the father and mother and the paternal and maternal ancestors of those ancestors subject to no limit, or to the limit imposed by seven and five degrees. If that was the object what was the use of stating with solemnity that there were three classes of bandhus'
72. I am clearly of opinion that the three classes of bandhus mean three classes alone and cannot he added to. The great (Sanskritscholar JogendraShiromani whose untimely death is still mourned, is of the same opinion and founds it, among other grounds, on the fact that the commentary and the text of the quotation do not use the same words: see Shiromani's ' Hindu Law, pp. 299 and 300.
73. Then the question is: how are the bandhus to be ascertained? We find that one's atmabandhus are to be found in one's father's line and in one's mother's line; for we find that father's sister's eon is an atmabandhu and he is descended from the grandfather of the propositos. Again We find that one's mother's sister's son and maternal uncle's son are atmabandhus, and they are the descendants of the maternal grandfather. These three instances lead us to the conclusion that one's atmabandhus are to be found among the descendants of one's paternal and maternal grandfathers.
74. In the case of pitribandhu we have three similar examples: father's father's sister's son is a pitribandhu and he is a descendant of father's grandfather. Again father's mother's sisters son and father's maternal uncle's son, who are one's bandhus, are to be found among the descendants of the father's maternal grandfather. We have therefore a further indication that for bandhus we have to search the lines of paternal great-grandfather and father's maternal grandfather. Similarly we find that mother's father's sister's son is a bandhu and he is to be found in the branch of mother's paternal grandfather. Again we have mother's mother's sister's and mother's maternal uncle's son, as bandhus, and as bandhus they are to be found in mother's maternal grandfather's line. If we put all these branches together, we find that four descriptions will suffice for tracing out the bandhus of a propositus. He the bandhu, may be one in the propositus's father's line which would include his own line, either ascending or descending because descending lines also would be his father's line through the propositus himself. Secondly, we shall find the bandhus in the propositus's maternal grandfather's line. Thirdly, we shall find a bandhu in propositus's father's maternal grandfather's line and fourthly, we shall find a bandhu in propositus's mother's maternal grandfather's line.
75. As pointed out by the learned author of the Mitakshara, when we talk of a 'line,' in dealing with Hindu law, we must understand that a line can only mean male ancestors and their descendants; for lineage is traced through the male ancestors and male descendants and nobody else. Mother's father's line would mean maternal grandfather, his father, his father and so on. Maternal grandfather's line or even mother's line would not mean mother's mother, her mother, her mother and so on. This view is supported by the rules by which cakes and oblations are offered to one's ancestors. A person offers cakes to his maternal grandfather, his father and to his father and not to the maternal grandmother and to her mother and to her mother. Sir Guru Das Banerji in his book on Marriage and Stridhan takes the same view: see his Book 4th Edn. p. 68. Putting together now all the rules that go to define a bandhu who has a right to inherit to a propositus, otherwise known as 'inheritable bandhu,' we find that he must possess the following qualifications: (a) He must belong to one of the four classes deduced by me from the statement of Vijnaneswar as to the bandhus being of three classes and from the illustration quoted by him. (b) The claim ant must be a sapinda coming within the seven and five degrees as described above. It will be noticed that many of the bandhus would not be a sapinda of the propositus; except for the definition that he (a bandhu) is a 'bhinna gotra sapinda' i.e., a sapinda who belongs to another family. Let us take this example:
76. In the above diagram, A is the common ancestor and 3 represents a son and D a daughter, S5 is not a sapinda of A, for A is not in the line of S5's mother, S5's mother's line would be the line of D3's father in which D2 is not, being only a mother. But S5 is a bandhu of A, inasmuch as A is in the line of the maternal grandfather of D3, the mother of S5, and is therefore a 'matribandhu' of S5. If S5 is a bandhu of A, it follows from the definition that S5 is a sapinda of A. It further follows that if two persona are not related with each other as sapindas in the ordinary way, but if they be each other's bandhu, they would thereby be the sapindaa of each, it again follows that the relationship of bandhu virtually expands the definition of sapinda in particular cases. In such cases we may say that S5 is a sapinda of A through his mother D3 in whose mother's line (paternal of course) A is.
77. We may now conclude that if the claimant to heirship be a bandhu of the deceased person but if the deceased person be not a sapinda of the claimant, the claim will fail, for Manu says: ' A sapinda's property shall be taken by a sapinda.' But if the deceased be related to the claimant as a bandhu, i.e., if both the deceased and the claimant be each other's bandhu, the claim will succeed.
78. So far I have examined the texts and have tried to come to my own conclusions. Let me now examine the decided cases. We find that their Lordships of the Privy Council in the case of Ram Chandra v. Vinaijak A.I.R. 1914 P.C. 1 laid down the teat of mutuality as to sapindaship and the test of sapindaship itself as two of the matters to be established for a successful claim as a bandhu. In the actual case before their Lordships of the Privy Council, the claim having failed on the ground that the claimants were claiming through their mother, and to succeed, must have been within five degrees of the common ancestor, Timaji, but were in, the sixth degree, it did become unnecessary for their Lordships of the Privy Council to consider whether the claimants were or were not bandhus. Nor was it necessary for their Lordships to consider what would have been the merits of the claim if the claimants had been claiming through their father and not through their mother. Their Lordships' general conclusions at p. 420 (of 42 Gal.) in Ram Chandra v. Vinayak A.I.R. 1914 P.C. 1, are based on the supposition that the claim is through the mother. This was pointed out by Rao, J., in Keshore Singh v. Secy. of State A.I.R. 1926 Mad. 881 at p. 690 (of 49 Mad.) and by me in the case of Harihar Prasad v. Ram Daur : AIR1925All17 .
79. A bandhu as we have seen must be related to the propositus, in a particular way, besides being a sapinda. In the case before their Lordships, the sapinda test failed, and it did not become necessary to go into the question of the particular relationships which came within the definition of bandhu. At pp. 403 and 404 (of 42 Cal.), their Lordships lay down the only issue before them, in the following language:
What we have to determine in this appeal is whether the term 'bandhu' has to be construed as the plaintiffs argued in the broadest sense or whether it is subject to any limitation. In the latter case what that limitation is according to the law by which the parties are governed?
80. Their Lordships found that one limitation to bandhuship was that imposed by the definition of 'sapindaship.' That test having sufficed for the failure of the plaintiffs, their Lordships made the following observations', (see p. 417 of 42 Cal.). These remarks-based on the case of Girdharilal v. Government of Bengal  12 M.I.A. 448 - hardly warrant the contention which is attempted to be based on them that the classes specified by Vijnaneswar can be added to:
In the present case however it does not seem necessary to their Lordships to enter upon the determination of the question whether the clauses can be extended, for the point at issue can ha decided on other grounds.
81. This case therefore is no authority for the proposition for which the learned Counsel for the appellants contends namely we have to see only one thing and one thing alone, namely, whether the claimant and propositus are or not within the prescribed degrees from the common ancestor. In other words the contention of the appellants' counsel is that if the propositus is related to the common ancestor through his father, we must see whether the propositus is or not within seven degrees from the common ancestor (including both the common ancestor and the propositus), and we have to see whether the propositus, if he is connected through his mother with the common ancestor, is within five degrees of the common ancestor or not. Similarly we have to apply the test in the case of the claimant. If the test of degrees' is satisfied, according to the learned Counsel for the appellants, the claimant would be entitled to succeed.
82. Let us now consider the facts of this present case. Sri Tewari the common ancestor of the propositus and of defendants 4 and 5 is not in the 'line' of the claimants' mother Kalawati. For although Sukhdeo is Kalawati's father, Mt. Badamo is not the father of Sukhdeo, but only his mother. Then let us seo if Kesho Prasad and Kali Prasad are ban-dhus of Sri Tiwari. Sri Tiwari is not in the direct father's line of the claimants nor is he in the line of maternal grandfather of them: he is therefore not an atmabandhu. He is not in the line of the claimants'father's maternal grand-father. He is not in the claimants' mother's maternal granefather's line. The result is that the claimants are not related to Sri Tiwari as bandhus and are therefore not bandhus of the deceased Suraj Praead, and are not his heirs. Coming to authorities again I find that a Full Bench of the Calcutta High Court in the case of Umaid Bahadur v. Udai Chand  6 Cal. 119 applied this test now applied by me : vide pp. 128 and 129.
83. The conclusion at which I have arrived is the same at which that eminent though much maligned modern writer Mr. Sarvadhikari arrived. People have not been able to follow his reasonings because he has put them in the shape of rules, as if they were framed by himself. But such is not really the case. He has tried to find some formula by which laymen can arrive at the right result in the question of a bandhu who may inherit. If we apply the tests deduced by me-from a consideration of the texts, we-shall find that the table of heirs framed, by Mr. Sarvadhikari arid adopted by all other important writers like Mayne, or' Hindu Law, is in accordance with my conclusion. I need hardly mention that; I am not considering the question of preference among the bandhus. That question does not arise in this case.
84. It will however be interesting to consider the method adopted by Mr. Sarvadhikari in finding out who a heritable bandhu is, and to find out the order of preference among them. Mr. Sarvadhikari lays down three rules, for finding out a heritable bandhu. The first two-rules cover, between them, a wide range of relations and the third rule excludes some of these. Those who fulfil all the three tests are, according to the learned' author, a bandhu These rules are to be found at pp. 597, 599, 601 and 602 ,Edn. 2 of the book,
85. No exception can be taken to the first rule. Let us take the case of the first ancestor. When a line passes through ' female the number of sapindas is five. As sapindaship has to be mutual, counting from a claimant who claims through a female, you can go only five degrees and descend to only five degrees, including the propositus, though the latter may be claiming entirely through a male line. In the case of ancestors the Nos. 2, 3 and 4 (see p. 597), the ascendants are counted through a female and therefore the limit has been again put at five degrees, i.e., at four excluding the propositus.
86. The second rule is equally unexceptionable. The father's 'bandhavas,' as mentioned in the text quoted by Vijnaneswar,, or 'connexions' being included among bandhus of the propositus, as also the 'connexions' of the mother, the sapindas. of these parents are likely to come in as bandhus of the propositus. But this inclusion of all the sapindas widens the-circle of relations too much. These may be sapindas but they are not necessarily bandhus. Bandhus must be in certain; lines of ancestors. Those that are not have to be excluded. Hence the third : rule. I have already pointed out that the definition of a bandhu, as we deduce its; from the rules framed by the author widens the list of sapindas strictly so' called. What Mr. Sarvadhikari has done is to widen the circle of the sapindas by including all the sapindas of the parents and then to exclude those who do not fulfil the teat of bandhuship.
87. I prefer the method of argument and drawing of the conclusion, mentioned by me, as the result of my own examination of the original texts. The result is the same as that arrived at by Mr. Sarvadhikari. As regards the method of finding out the comparative merits of the bandhus for the purpose of preference of inheritance, the question does not arise in this case, and I need not express any opinion. I may however mention what according to my reading Mr. Sarvadhikari has done only for the purpose of finding out the nearness of relationship for it is the nearness that gives the claim to inheritance. Mr. Sarvadhikari has divided the bandhus into (1) at ma band bus ex parte paterna; (2). at ma bandhus ex parte materna (3) pitri bindhus, all ex parte materna (4) matri bandhus, all ex parte materna.
88. This method has not been adopted to contradict Vijnaneswar or to improve on his method, but it is in keeping with his method. The persons who claim through the father in the paternal line or through the mother in the mother's paternal line have all been classed as atma bandhus. Those that claim through the father's mother have been classed as pitri-bandhus, and those that claim through the mother of the mother have boon classed as matribandhus. As I have sail, this has been done solely for the purpose of finding out the nearness of relationship. This method does not in any way clash with the opinion of Vijnaneswar that the atama bandhus (as described by him, by means of the quotation) take first and then pitribandhus and then the matribandhus.
89. However all these matters do not arise for discussion in this case, my object in entering into the discussion was to show, that my conclusions, being the same as that of Mr. Sarvadhikari, receive corroboration from him and that corroboration by Mr. Sarvadhikari is not to be discounted on the ground that Mr. Sarvadhikari is full of mistake's or that he has laid down arbitrary rules. I may mention that this division of the bandhus for the purpose of finding out the nearness has been adopted by Jogendra Shiromani also. It now remains for me to consider the few cases cited on behalf of the-claimants to bandhuship, the appellants. The case of Kesar Singh v. Secy. of State A.I.R. 1926 Mad. 881 entirely supports the appellants' case. But with all respect I have not been able to accept it as having been correctly decided.
90. Spencer, J., based his decision on several considerations. One of these is that; the word 'santan' which has been, translated as a 'line' by all 'translators, does not mean male ascendants alone 'Santana,' no doubt has another meaning, viz. issue, i.e. either male or female. But it also means a line.' Vijnaneswar himself interprets it as meaning male line alone. I have quoted his very words. Then the learned Judge? brushes, aside the case of Umaid Bahadur  6 Cal. 119 as an obiter dictum. But it has already been cited with approval in the Privy Council case of Ram Chandra A.I.R. 1914 P.C. 1. Then, he appears to have accepted an extension of the classes of bandhus. Rao, J., busied., himself with an examination of the reasonings (for his conclusions) of Mr. Sarvadhikari, and did not examine the: texts himself.
91. The next case that was cited was the case Umashankar v. Nageswar  3 Pat. L.J. 663. The observations of Malik, J., were relied on. where he examined the book by Mr.. Sarvadhikari. The point before the Patna Court was one of relative title of two claimants, both of whom were bandhus. But the question before us-did not at all arise in the case. The question was which of the two claimants-had the better title. I am not concerned with the defence of Mr. Sarvadhikari. He can take care of himself. But when, all is said, it would still appear that - questions of precedence apart the table of bandhus prepared by him and adopted by other writers, is substantially correct. I am using the word 'substantially'' because this case is not concerned with each of the numerous cases of bandhus shown in the tables, and I have no. examined each of them separately. Excluding the cases in 49 Mad. and 6 Cal. there is no other case directly in point. As the result of my own reading of the original Mitakshara and relevant texts, and on the authority of the case in 6 Cal. I have arrived at the conclusion that the Madras case does not lay down the correct law. My answer to the question before the Pull Bench is that Kesho Prasad and Kali Prasad are not heritable bandhus of Suraj Prasad.
92. I concur and have nothing to add.