Cover of the Madhyama Āgama (Middle-Length Discourses), Volume I
BDK America
Publish Date: 
Marcus Bingenheimer, Editor in Chief; Bhikkhu Anālayo and Roderick S. Bucknell, Co-Editors, et al.


9.5 x 6.6 x 1.8
$80.00 + shipping

PDF Download



The Madhyama Āgama, or Middle-length Discourses, belong to the earliest stratum of Buddhist literature. Here are dialogues between Śākyamuni Buddha and his disciples and followers, covering such topics as Karma, death and rebirth, and the training of the mind required to reach nirvana. First volume of a projected three-volume set.

Includes content presented in English for the first time ever.


The scope of this project is suggested by the size of the team of editors and translators who have brought it to fruition. They include Marcus Bingenheimer, Editor in Chief; Bhikkhu Anālayo and Roderick S. Bucknell, Co-Editors; and translators Kin-Tung Yit, William Chu, Teng Weijen, Shi Chunyin, and Kuan Tse-fu. Bhikkhu Anālayo and Marcus Bingenheimer also contributed direct translations of several sections. See the Table of Contents to see which translator prepared divisions of the text. 

Taishō 26

Volume 1

The Madhyama Āgama (Middle-Length Discourses)

The Madhayama Āgama, or Middle Length Discourses, belong to the earliest stratum of Buddhist sutra literature. The dialogues between Śākyamuni Buddha and his disciples and followers present the teachings of early Buddhism on karma, death and rebirth, and the training of the mind in order to reach nirvana. The speakers often make use of parables and tales to illustrate the correct understanding of early doctrine.

The Madhayama Āgama was translated into Chinese from a now lost Indian original in 398 C.E. It corresponds mainly to the Majjhima Nikāya of the Pali canon, although many of its sutras have parallels in other sections of the Pali canon.

Volume I contains the translations of divisions 1-6 (of 18) and sutras 1-71 (of 222). Translations of the remaining divisions and sutras will be published in two subsequent volumes.

Skt. Madhyamāgama, translated by Gautama Saṃghadeva into the Chinese as Zhong ahan jing (長阿含經). 60 fascicles.

Table of Contents

A Message on the Publication of the English Tripiṭaka     NUMATA Yehan    v
Editorial Foreword    MAYEDA Sengaku    vii
Publisher’s Foreword A.    Charles Muller    ix
Introduction    Marcus Bingenheimer    xv
Abbreviations    xxviii

The Madhayama Āgama (Middle-length Discourses), Volume I

Division 1. On Sets of Seven
Translated by Bhikkhu Anālayo

1. The Discourse on Wholesome Qualities 3
2. The Discourse on the Coral Tree 8
3. The Discourse with the Parable of the [Border] Town 10
4. The Discourse with the Water Parable 17
5. The Discourse with the Parable of the Heap of Wood 23
6. The Discourse on the Destination of a Good Person 33
7. The Discourse on [Sources of] Worldly Merit 38
8. The Discourse on Seven Suns 41
9. The Discourse on Seven Chariots 47
10. The Discourse on the Cessation of Taints 56

Division 2. On Karma
Translated by Kin-Tung Yit

11. The Discourse with the Parable [of the Ounce] of Salt 63
12. The Discourse to Vappa 68
13. The Discourse on [Tenets to Be] Transcended 73
14. The Discourse to Rāhula 78
15. The Discourse on Intention 84
16. The Discourse to the Kālāmas 89
17. The Discourse to Gāmaṇi 96
18. The Discourse to Sīha 101
19. The Discourse to the Nigaṇṭhas 108
20. The Discourse to Pāṭaliya 122

Division 3. Sāriputta
Translated by William Chu

21. The Discourse on an Even Mind 141
22. The Discourse on Perfecting the Precepts 146
23. The Discourse on Wisdom 152
24. The Discourse on the “Lion’s Roar” of Sāriputta 160
25. The Discourse with Parables Relating to Water 166
26. The Discourse to the Gulissāni 170
27. The Discourse to Dhānanjāni 176
28. The Discourse on Teaching the Ill [Anāthapiṇḍika] 188
29. The Discourse by Mahā Koṭṭhita 201
30. The Discourse with the Parable of the Elephant’s Footprint 218
31. The Discourse on Discerning the Noble Truths 234

Division 4. On Extraordinary Qualities
Translated by Teng Weijen

32. The Discourse on Extraordinary Qualities [of the Buddha] 247
33. The Discourse on [How Ānanda Became] the Attendant 257
34. The Discourse by Bakkula 271
35. The Discourse to the Asura 275
36. The Discourse on Earthquakes 281
37. The Discourse at Campā 286
38. The [First] Discourse to the Householder Ugga 292
39. The [Second] Discourse to the Householder Ugga 300
40. The Discourse to the Householder Hatthaka 306
41. The Discourse on the Householder Hatthaka 316

Division 5. On Conditions
Translated by Marcus Bingenheimer

42. The Discourse on “What is the Purpose?” 319
43. The Discourse on No [Need for] Thought 321
44. The Discourse on Mindfulness 322
45. The [First] Discourse on Shame and Scruples 323
46. The [Second] Discourse on Shame and Scruples 324
47. The [First] Discourse on the Precepts 326
48. The [Second] Discourse on the Precepts 326
49. The [First] Discourse on Respect 327
50. The [Second] Discourse on Respect 328
51. The Discourse on the Beginning 329
52. The [First] Discourse on Nutriments 332
53. The [Second] Discourse on Nutriments 339
54. The Discourse on [Attaining the] Wisdom of Cessation [of the Taints] 343
55. The Discourse on Nirvana 346
56. The Discourse to Meghiya 349
57. The Discourse Spoken for the Monks 354

Division 6. On Kings
Translated by Shi Chunyin (Fascicles 11–13) and Kuan Tse-fu (Fascicles 14–16)

58. The Discourse on the Seven Treasures 359
59. The Discourse on the Thirty-two Marks 360
60. The Discourse on the Four Continents 365
61. The Discourse with the Cow Dung Parable 373
62. The Discourse on King Bimbisāra Meeting the Buddha 379
63. The Discourse at Vebhaḷiṅga 387
64. The Discourse on the Divine Messengers 407
65. The Discourse with the Raven Parable 422
66. The Discourse on Origins 432
67. The Discourse on Mahādeva’s Mango Grove 447
68. The Discourse on Mahāsudassana 462
69. The Discourse with the Thirty Analogies 477
70. The Discourse on the Wheel-turning Monarch 483
71. The Discourse to Pāyāsi 505

Notes 541
Bibliography 557
Index 565
A List of the Volumes of the BDK English Tripiṭaka (First Series) 589


From the Introduction

by Marcus Bingenheimer, Editor in Chief

To translate is to explain.1

—Sengyou (445–518)

The Zhong ahan jing (Skt. Madhyamāgama, T. 26), which is translated here, is one of the four major canonical collections of early Buddhist sutras preserved in Chinese.2 The sutras grouped in this collection were deemed to be of “middle length” (Ch. zhong, Skt. madhyama: “middle”). The Chinese Ahan (Skt. Āgama) collections correspond to the better-known Nikāyas of the Pāli canon. While the latter texts (the Dīghanikāya, Majjhimanikāya, Saṃyuttanikāya, and Aṅguttaranikāya) have long been available in English translation, none of the equally important Āgamas has so far been rendered into English, or indeed any Western language.3 The Āgamas and Nikāyas constitute the Sūtra-piṭaka4 for the northern and southern transmission5 of Buddhism respectively, and they are the primary sources for what we know about early Buddhist doctrine. They remain of immense importance for the study and practice of Buddhism in both academic and religious contexts. With this translation we aim to make accessible another important witness, a view back onto the earliest period of the Buddhist tradition....

[Web Editor's Note: In this introduction, Mr. Bingenheimer, the Editor in Chief for the present text, outlines its development “from its beginning in the oral tradition, through its Indian and Chinese ‘incarnations,’ up to the present English translation.” The excerpt that follows demonstrates the project's rigor and sets the stage for understanding the multiplicity of texts and sources in the effort.]

It is not yet possible to give a comprehensive account of how exactly the Āgamas relate to the Nikāya corpus; the general outline, however, is clear. The four Āgamas, all of which are now preserved intact only in Chinese, represent portions of the Sūtra-piṭakas of various northern schools of Buddhism, which were transmitted in various Indian dialects and then made their way to the Buddhist kingdoms of Central Asia and along the Silk Road into China. The Nikāyas contain the sutra texts of the southern tradition that were transmitted to Sri Lanka and preserved in the Pāli canon, whose formative period was from the first century B.C.E. to the fifth century C.E.

Regarding the relationship between the Zhong ahan jing and the Majjhimanikāya, Minh Chau’s groundbreaking 1964 study, The Chinese Madhyama Āgama and the Pāli Majjhima Nikāya (Delhi: Motilal Banarsidass, 1991, reprint), describes a number of differences in doctrinal and narrative content.6 Bhikkhu Anālayo’s A Comparative Study of the Majjhima-nikāya (Taipei: Dharma Drum Publishing, 2011) discusses in great detail all Chinese parallels to the sutras of the Majjhimanikāya, many of which are represented in our translation. In the course of his research Bhikkhu Anālayo has also translated and discussed a large number of sutras from the Zhong ahan jing. In view of the extent of his contribution, the relevant citations are listed in an addendum to the bibliography.

As the translation work for this volume has shown again and again, comparison with the Pāli versions is often essential for an understanding of difficult passages. Although most Āgama sutras preserved in Chinese do have counterparts in the Pāli canon and vice versa, they are frequently found in different collections. According to Akanuma Chizen’s catalogue, Kanpa shibu shiagon goshōroku—The Comparative Catalogue of Chinese Āgamas and Pāli Nikāyas (Nagoya: Hajinkakushobō, 1929), of the two hundred and twentytwo sutras of T. 26, only one hundred and three have their counterpart in the Majjhimanikāya; fourteen have their counterpart in the Dīghanikāya, seventeen in the Saṃyuttanikāya, and eighty-seven in the Aṅguttaranikāya.7 Fourteen of the two hundred and twenty-two sutras of T. 26 have no known parallel in the Pāli corpus.


After the death of the founder, Buddhist texts were transmitted orally in Middle Indo-Aryan dialects (Prakrits). While the southern tradition eventually settled on one of these dialects, Pāli, as its canonical language, in India and Central Asia Buddhist texts were successively Sanskritized and/or translated into other languages such as Chinese, Tokharian, Khotanese, Sogdian, and Tibetan.8 Also, new Buddhist texts in India, from at least the third century onward, were directly composed in standard Sanskrit. Manuscripts from the northern tradition, especially those of Central Asian provenance, are therefore often in Prakrit (especially Gāndhārī)9 or some nonstandard form of Sanskrit, sometimes called Buddhist Sanskrit, an intermediate stage between some Prakrit and standard Sanskrit.

Unfortunately, we do not have a complete Madhyamāgama in either Sanskrit or Prakrit. Only for the Dīrghāgama do we have a (largely) complete Indic manuscript;10 for the other three collections there are only numerous manuscript fragments containing sometimes one or more sutras but more commonly only a few lines of text. These fragments stem from different periods and locations. As yet there is no combined edition of this material for any Āgama, nor is there general bibliographic overview. For the Madhyamāgama, however, Jin-il Chung and Takamichi Fukita have catalogued all available Sanskrit fragments in A Survey of the Sanskrit Fragments Corresponding to the Chinese Madhyamagama, Including References to Sanskrit Parallels, Citations, Numerical Categories of Doctrinal Concepts, and Stock Phrases. (Tokyo: Sankibo, 2011).11

As with most other long texts that have been transmitted orally, and since the first century B.C.E. with the support of writing, we have to assume that no two instances of a text were ever fully identical, even though the Indian oral tradition is known for preserving texts over long periods with very little variation.12 If nothing else, the fact that the Madhyamāgama and the Majjhimanikāya share only about one hundred sutras out of two hundred and twentytwo (MĀ) and one hundred and fifty-two (MN), shows that the Āgamas/ Nikāyas as collections were still in flux for a considerable time. In their current form the collections can be attested to the fourth to fifth centuries. Only for the period since the translation into Chinese in the north (fourth century C.E.) and Buddhaghosa’s commentaries in Sri Lanka (fifth century C.E.) can we be confident that the gestalt of the four major sutra collections has remained more or less unchanged. Generally, it is only after that time that it is possible to date further changes in either tradition.13

The Sanskritized texts circulating in Central Asia continued to evolve after the fifth century, as is evident from the differences between later Sanskrit remains and the Chinese versions, but owing to the fragmentary nature of the Sanskrit materials our knowledge about these changes is extremely limited.14

The picture that emerges when the Sanskrit fragments are compared with the Chinese and Pāli versions is that several versions of the texts collected in the Āgamas/Nikāyas did exist between the first century B.C.E. and the fourth century C.E. During this time, when the oral tradition began to be supported by writing in both the northern and southern traditions, each document instance would have differed slightly, reflecting school affiliation, local traditions, linguistic environment, nonstandardized scripts, or any combination of these factors. Moreover, the manuscript evidence shows that some sutras contained in the Āgamas/Nikāyas circulated independently as well, a fact that is also borne out by the many instances of individual translations of Āgama/Nikāya sutras available in Chinese.

To illustrate the situation with one example: the Shansheng jing (sutra 135 of T. 26) exists in a cluster with at least thirteen other instances of the text: the Siṅgālovāda-sutta of the Dīghanikāya (DN, no. 31), the Shansheng jing of the Chinese Dīrghāgama (T. 1, no.16), the various Sanskrit fragments of the Śikhālaka-sūtra, two other individual Chinese translations (T. 16 and T. 17), and a quotation in the Mahākarma-vibhaṅga.15 There exist complex stemmatic relationships among these texts, which—as for most sutra clusters— have not yet been clarified and perhaps will never be.

We are therefore faced with the fact that the collection of sutras translated in this volume is but one witness that has survived and attained prominence, like a snapshot of the development of the text taken at a certain time at a certain place. The composition of the text is therefore to a degree arbitrary, in the sense that T. 26 would in all likelihood have looked rather different if the Chinese translation had been done earlier, or later, or from a version of another school. In this light, the statement “the Majjhimanikāya has one hundred and fifty-two sutras and the Madhyamāgama two hundred and twenty-two” loses some of its apodictic edge. It so happened that the Majjhimanikāya found closure with one hundred and fifty-two sutras, while incidentally the only surviving Madhyamāgama (preserved in Chinese as T. 26) found closure with two hundred and twenty-two sutras.

This angle is emphasized here because all too often doctrinal statements are made on the basis of a single passage in a single instance of a text. To remember that some randomness is part of tradition does not amount to relativism or to the belief that there is no message at all in these texts; it means, rather, that there is no final authoritative foundation on which to build one’s argument beyond the multiplicity and careful navigation of the text-clusters....



1. T.55.2145:4c5–6. All references in this format were retrieved from and are based on the CBETA corpus (CD Version 2009), which contains a digital edition of the Taishō Canon.

2. The three others are the Chang ahan jing (Dīrghāgama, T. 1), the Za ahan jing (Saṃ - yuk tāgama, T. 99), and the Zengyi ahan jing (Ekottarikāgama, T. 125).

3. All Chinese Āgamas have been translated into Japanese and Korean.

4. In the Pāli canon the Sutta-piṭaka includes a fifth collection of texts, the Khudda - kanikāya. The schools of the northern tradition knew a similar group of texts, but some established it as a separate piṭaka, the Kṣudraka-piṭaka, instead of including it in the Sūtra-piṭaka (though the designation Kṣudrakāgama is also found in some sources: Egaku Mayeda, “Japanese Studies on the Schools of the Chinese Āgamas,” in Heinz Bechert, ed., Zur Schulzugehörigkeit von Werken der Hīnayāna-Literatur [Gottingen: Vandenhoeck & Ruprecht, 1985], p. 95; Etienne Lamotte, Le Traité de la Grande Vertu de Sagesse de Nāgārjuna, vol. 3 [Louvain: Universite de Louvain Institut Orientaliste, 1970]). Piṭaka or Āgama, the fifth collection was never translated as such into Chinese.

5. For want of better terms we subsume under the term “northern tradition” the textual traditions of the schools in Northern India and Central Asia that were codified in languages other than Pāli; the textual traditions that were transmitted to Sri Lanka and eventually resulted in the Pāli canon are correspondingly termed “southern.” This distinction may be blunt, but is nevertheless useful.

6. See Bhikkhu Anālayo, “The Chinese Madhyama-āgama and the Pāli Majjhimanikāya— In the Footsteps of Thich Minh Chau,” Indian International Journal of Buddhist Studies 9 (2008): 1–21.

7. The figures cited here are query results from the digital version of Chizen Akanuma’s Kan pa shibu shiagon goshōroku (comcatV3.xml, available at http://mbingenheimer. net/tools/comcat/indexComcat.html). Added together, these numbers slightly exceed the total of two hundred and twenty-two, because in some cases one sutra has more than one parallel. Moreover, these numbers are subject to debate. Bhikkhu Anālayo, in his detailed study, A Comparative Study of the Majjhima-nikāya (Taipei: Dharma Drum, 2011, p. 9), allows for only ninety-six parallels of the Zhong ahan jing in the MN. While the absolute numbers will always depend on the definition of Madhyama Agama Vol. 1 REPRINT_Layout 1 6/18/14 12:52 PM Page 541 “parallel,” this general ratio will hold true: less than half of the sutras of T. 26 are found in the MN, most of the rest are found in the three other Nikāyas, and for a small percentage there seems to be no clear parallel in the Pāli corpus. Akanuma’s catalogue, Kanpa shibu shiagon goshōroku (The Comparative Catalogue of Chinese Āgamas and Pāli Nikāyas) (Nagoya: Hajinkaku shobō, 1929) is now dated; for the MN parallels one should refer to Bhikkhu Anālayo and Roderick S. Bucknell, “Correspondence Table for Parallels to the Discourses of the Majjhima Nikāya: Toward a Revision of Akanuma’s Comparative Catalogue,” Journal of the Centre for Buddhist Studies, Sri Lanka 4 (2006): 215–243, which has also been included in a recent attempt to synthesize the available data in database format (

8. Only a few Āgama sutras were translated into Tibetan. For an in-depth discussion of these, see Peter Skilling, “Theravādin Literature in Tibetan Translation,” Journal of the Pali Text Society 19 (1993): 69–203.

9. See Oskar von Hinuber, “Origin and Varieties of Buddhist Sanskrit,” in Colette Caillat, ed., Dialectes dans les Littératures Indo-aryennes (Paris: de Boccard, 1989), pp. 341–367, on the varieties of Buddhist Sanskrit. See Richard Salomon, Ancient Buddhist Scrolls from Gandhāra—The British Library Kharoṣṭī Fragments (Seattle: University of Washington Press, 1999), pp. 110–140, for an overview of Gāndhārī and evidence that Gāndhārī was indeed one single Prakritic language that exerted a strong influence on the transmission of Buddhist texts. It has been argued that the source texts of the Chang ahan jing (Dīrghāgama, T. 1) (Ernst Waldschmidt, “Remarks on the Madhyamāgama Ms. Cat.-no.412,” in Ernst Waldschmidt, Walter Clawiter, Lore Sander, and Preussische Turfan-Expeditionen, eds., Sanskrithandschriften aus den Turfanfunden, vol. IV, pp. 1–5 [Wiesbaden: F. Steiner, 1980]), and the Pu yao jing (Lalitavistara, T. 186) (John Brough, “The Arapacana Syllabary on the Old Lalita - vistara,” Bulletin of the School of Oriental and African Studies 40 [1977]: 85–95) were in Gāndhārī. On the language of the source text for the Chang ahan jing (T. 1), see Seishi Karashima, Chōagonkyō no gengo no kenkyū (A Study of the Original Language of the Chinese Dīrghaāgama) (Tokyo: Hirakawa shuppan, 1994). Daniel Boucher, in “Gāndhārī and the Early Chinese Buddhist Translations Reconsidered: The Case of the Saddharmapuṇḍarīkasūtra,The Journal of the American Oriental Society 118 (4) (October 1998): 471–506, has criticized the trend to assume by default that Gāndhārī was the original language of most of the early translations, and emphasizes that other factors connected to the orality of the translation process must be taken into account when reconstructing the Indian source text from the Chinese.

10. Jens-Uwe Hartmann, “Further Remarks on the New Manuscript of the Dīrgha-āgama,” Jour nal of the International College for Advanced Buddhist Studies 5 (2002): 133–150 (98–81); and “Contents and Structure of the Dīrghāgama of the (Mūla-) Sarvāstivādins,” Annual Report of the International Research Institute for Advanced Buddhology 7 (2004): 119–137.Various parts of this manuscript have been edited: Oliver von Criegern, Das Kūṭatāṇḍyasūtra. Nach dem Dīrghāgama-Manuskript herausgegeben und übersetzt, unpublished MA thesis, Ludwig-Maximilians-Universitat Munchen, 2002; Gudrun Melzer, Ein Abschnitt aus dem Dīrghāgama, Teil 1, unpublished Ph.D. Madhyama Agama Vol. 1 REPRINT_Layout 1 6/18/14 12:52 PM Page 542 dissertation, Ludwig-Maximilians-Universitat, Munchen, 2006; Lita Peipina, The Piṅ- galātreya sūtra of the (Mūla)sarvāstivādins: Its Edition and Study. Investigation of the Piṅgalātreya sūtra’s Status within the Dīrghāgama Collection of “Long Discourses of the Buddha,” unpublished MA thesis, Department of Culture Studies and Oriental Languages, University of Oslo, 2008; Lixiang Zhang, Das Śaṃkarasūtra: Eine Übersetzung des Sanskrit-Textes im Vergleich mit der Pāli Fassung, unpublished MA thesis, Ludwig-Maximilians-Universitat, Munchen, 2004; Chunyang Zhou, “Das Kaivartisutra der neuentdeckten Dirghāgama-Handschrift: Eine Edition und Rekonstruktion des Textes,” unpublished MA thesis, Gottingen, 2008.

11. Anālayo “Zhong Ahan,” in W. G. Weeraratne, ed., Encyclopaedia of Buddhism (Sri Lanka: Department of Buddhist Affairs, 2009), vol. 8, no. 3, pp. 827–830, includes a helpful list of Sanskrit fragments published in the first nine volumes of the ten-volume Sanskrithandschriften aus den Turfanfunden, Ernst Waldschmidt, et al. (Wiesbaden: F. Steiner, 1965–2004).

For the Chinese Saṃyuktāgama translations (T. 99, T. 100, T. 101) Jin-il Chung, A Survey of the Sanskrit Fragments Corresponding to the Chinese Saṃyuktāgama (Tokyo: Sankibo, 2008), provides bibliographic information for all known Sanskrit parallels to the Chinese Saṃyuktāgama. Fumio Enomoto, A Comprehensive Study of the Chinese Saṃyuktāgama: Indic Texts Corresponding to the Chinese Saṃyuktāgama as found in the Sarvāstivāda-Mūlasarvāstivāda Literature (Kyoto: Kacho Junior College, 1994), collates the known Sanskrit fragments corresponding to the Saṅgītanipāta (Sagāthavagga) of the Chinese Saṃyuktāgama.

Numerous quotations from the Sanskrit Āgamas can be found in the Abhidharma - kośa and (in Tibetan translation) in one of its commentaries, Śamathadeva’s Abhi - dharmakośopāyikā.Ma dhyamāgama passages in the former can be located in Bhikkhu Pāsādika, Kanonische Zitate im Abhidharma kośabhāṣya des Vasubandhu (Gottingen: Vandenhoeck & Ruprecht, 1989), especially p. 135; for Madhyamāgama passages in the latter, see Yoshifumi Honjō, A Table of Āgama Citations in the Abhi - dharmakośa and the Abhidharmakośopāyikā (Kyoto: Privately published, 1984). The Āgama quotations in the Upāyikā are being translated by Sāmaṇerī Dhammadinnā. The first installment was published as “A Translation of the Quotations in Śamatha - deva’s Abhidharmakośopāyikā-ṭīkā Parallel to the Chinese Saṃyukta-āgama Discourses 8, 9, 11, 12, 17 and 28,” Dharma Drum Journal of Buddhist Studies 11 (2012): 63–96. 

12. See Bhikkhu Anālayo, “Oral Dimensions of Pali Discourses: Pericopes, Other Mnemonic Techniques and the Oral Performance Context,” Canadian Journal of Buddhist Studies 3 (2007): 5–33, on the reliability of transmission in the Pāli tradition and further references.

13. For instance, the confusion of the sutra order in the longer Chinese Saṃyuktāgama (T. 99) before the ninth century (Roderick S. Bucknell, “The Structure of the Sagātha- Vagga of the Saṃyutta-Nikāya,Buddhist Studies Review 24 [1] [2007]: 7–34), or the forking of the shorter Chinese Saṃyuktāgama (T. 100) into two versions (Bucknell, “The Two Versions of the Other Translation of Saṃyuktāgama,Chung-Hwa Journal of Buddhist Studies 21 [2008]: 23–54) in or before the twelfth century.

14. Cf. the differences between some passages in T. 26 and later Sanskrit texts pointed out by P. V. Bapat, “Chinese Madhyamāgama and the Language of its Basic Text,” in B. P. Sinha, ed., Dr. Satkari Mookerji Felicitation Volume (Varanasi: Chowkhamba Publications, 1969), p. 2. See Waldschmidt, “Remarks on the Madhyamāgama Ms. Cat.-no. 412,” and Lore Sander, “Fixed Sequences of Texts in some Sūtra Collections,” in Waldschmidt, et al., eds., Sanskrithandschriften aus den Turfanfunden, vol. IV, pp. 6–12, for examples of how at least some sutra sequences contained in the Madhyamāgama were included in fragments of other, as yet unidentified collections as well.

15. Jens-Uwe Hartmann and Klaus Wille, “A Version of the Śikhālakasūtra/Siṅgālovādasutta,” in Jens Braarvig, Paul Harrison, Jens-Uwe Hartmann, Kazunobu Matsuda, and Lore Sander, eds., Manuscripts of the Schoyen Collection (Oslo: Hermes Academic Publishing, 2006), vol. 3, pp. 1–6.