APPENDIX III - DRAFT REVISED RECOMMENDED METHODS OF SAMPLING FOR THE DETERMINATION OF PESTICIDE RESIDUES FOR COMPLIANCE WITH MRLS (Advanced to Step 8 of the Codex Procedure)

1. OBJECTIVE
2. PRINCIPLES
3. SAMPLING PROCEDURES
4. CRITERIA FOR DETERMINING COMPLIANCE
Annex I. DEFINITION OF TERMS
Annex II. SCHEMATIC REPRESENTATION OF SAMPLING
REFERENCES

1. OBJECTIVE

The objective of these sampling procedures is to enable a representative sample to be obtained from a lot, for analysis to determine compliance with Codex Maximum Residue Limits (MRLs).

2. PRINCIPLES

2.1 Codex MRLs are intended to ensure good agricultural practices in the use of pesticides and are set at the appropriate levels required to minimize exposure of consumers and animals and to protect crops, food or feeding stuffs.

2.2 A Codex MRL for a plant, egg or dairy product takes into account the maximum level expected to occur in a composite sample, which has been derived from multiple units of the treated product and which is intended to represent the average residue level in a lot. A Codex MRL for meat and other poultry products takes into account the maximum level expected to occur in the tissues of individual treated animals or birds.

2.3 In consequence, MRLs for meat and poultry products apply to a bulk sample derived from a single primary sample, whereas MRLs for plant products, eggs and dairy products apply to a composite bulk sample derived from 1-10 primary samples.

3. SAMPLING PROCEDURES

Notes.

(a) The terms used are defined in Annex I and the procedures are shown schematically in Annex II.

(b) ISO recommendations for sampling of grain¹, or other commodities shipped in bulk may be adopted, if required.

Contamination and deterioration of samples must be prevented at all stages, because they may affect the analytical results. Each lot to be checked for compliance must be sampled separately.

3.2 Collection of primary samples

The minimum number of primary samples to be taken from a lot is determined from Table 1. Each primary sample should be taken from a randomly chosen position in the lot, as far as practicable. The primary samples must consist of sufficient material to provide the laboratory sample(s) required from the lot.

Notes. (a) Sampling devices required for grain¹, pulses² and tea³ are described in ISO recommendations and those required for dairy products⁴ are described by the IDF.

3.3.1 Procedure for meat and poultry products (Table 3)

Each primary sample is considered to be a separate bulk sample and it should be mixed well, if practicable.

3.3.2 Procedure for plant products, eggs or dairy products (Tables 4 and 5)

The primary samples should be combined and mixed well, if practicable, to form the bulk sample.

3.3.3 Alternative procedure where mixing to form the bulk sample is inappropriate or impractical

Where units may be damaged (and thus residues may be affected) by the processes of mixing or sub-division of the bulk sample, or where large units cannot be mixed to produce a more uniform residue distribution, the units should be allocated randomly to replicate laboratory samples at the time of taking the primary samples. In this case, the bulk sample is considered to be the sum of the laboratory samples analyzed.

3.4 Preparation of the laboratory sample

Where the bulk sample is larger than is required for a laboratory sample, it should be divided to provide a representative portion. A sampling device, quartering, or other appropriate size reduction process may be used but units of fresh plant products or whole eggs should not be cut or broken. Where required, replicate laboratory samples should be withdrawn at this stage or they may be prepared as in 3.3.3, above. The minimum sizes required for laboratory samples are given in Tables 3 and 4.

3.5 Sampling record

The sampling officer must record the nature and origin of the lot; the owner, supplier or carrier of it; the date and place of sampling; and any other relevant information. Any departure from the recommended method of sampling must be recorded. A signed copy of the record must

accompany each replicate laboratory sample and a copy should be retained by the sampling officer.

3.6 Packaging and transmission of the laboratory sample

The laboratory sample must be placed in a clean, inert container which provides secure protection from contamination, damage and leakage. The container should be sealed, the sampling record must be attached and the sample delivered to the laboratory as soon as practicable. Spoilage in transit must be avoided, e.g. fresh samples should be kept cool and frozen samples must remain frozen. Samples of meat and poultry products should be frozen prior to despatch, unless transported to the laboratory before spoilage can occur.

3.7 Preparation of the analytical sample

The laboratory sample should be given a unique identifier which, together with the date of receipt and the sample size, should be added to the sample record. The part of the commodity to be analysed^5,6, i.e. the analytical sample, should be separated as soon as practicable. Where the residue level must be calculated to include parts which are not analysed, the weights of the separated parts must be recorded.

3.8 Preparation and storage of the analytical portion

The analytical sample should be comminuted, if appropriate, and mixed well, to enable representative analytical portions to be withdrawn. The size of the analytical portion should be determined by the analytical method and the efficiency of mixing. The methods for comminution and mixing should not affect the residues present in the analytical sample. Where appropriate, the analytical sample should be processed under special conditions, e.g. at sub-zero temperature, to minimize adverse effects. Where processing could affect residues and where practical alternative procedures are not available, the analytical portion may consist of whole units, or segments removed from whole units. If the analytical portion thus consists of few units or segments, it is unlikely to be representative of the analytical sample and sufficient replicate portions must be analysed, to indicate the uncertainty of the mean value. If analytical portions are to be stored before analysis, the method and length of time of storage should be such that they do not affect the level of residues present. Additional portions must be withdrawn for replicate and confirmatory analyses, as required.

4. CRITERIA FOR DETERMINING COMPLIANCE

Table 1. Minimum number of primary samples to be taken from a lot
Table 2. Number of randomly selected primary samples required for a given probability of detecting at least one non-compliance in a lot of meat or poultry product
Table 3. Meat and poultry products: Description of primary samples and minimum size of laboratory samples
Table 4. Plant products: Description of primary samples and minimum size of laboratory samples
Table 5. Egg and dairy products: Description of primary samples and minimum size of laboratory samples

4.1 Analytical results must be derived from samples which were in a fit state for analysis and they must be supported by acceptable quality control data (e.g. for instrument calibration and pesticide recovery - refer to Codex Alimentarius, Volume 2, Section 4.2, “Guidelines on good laboratory practice in pesticide residue analysis”). Results should not be corrected for recovery. Where a residue is found to exceed an MRL, its identity should be confirmed and its concentration must be verified by analysis of one or more additional analytical portions.

4.2 The Codex MRL applies to the bulk sample.

4.3 The lot complies with a Codex MRL where the MRL is not exceeded by the analytical result(s).

4.4 Where results for the bulk sample exceed the MRL, a decision that the lot is non-compliant must take into account: (i) the range of results obtained from replicate laboratory samples and/or replicate analytical portions, as applicable; and (ii) the accuracy and precision of analysis, as indicated by the supporting quality control data.

Table 1. Minimum number of primary samples to be taken from a lot

			Minimum number of primary samples to be taken from the lot
(a) Meat and poultry products
	a non-suspect lot		1
	a suspect lot		approximately 6-30	(see note(i) , below)
(b) Plant products, eggs and dairy products
	(i) Products, packaged or in bulk, which can be assumed to be well mixed or homogeneous		1	see note (d) under definition of a lot, Annex 1
	(ii) Products, packaged or in bulk, which may not be well mixed or homogeneous			see note (ii), below
	either:
		Weight of lot, kg
		<50	3
		50-500	5
		>500	10
	or
		Number of cans, cartons or other containers
		in the lot
		1-25	1
		26-100	5
		>100	10

Notes.

(i) If the location of contaminated units within a lot of a meat, dairy or poultry product cannot be determined by visual inspection, the number of samples to be taken from a suspect lot will depend on the degree of confidence required (see Table 2).

(ii) For products comprised of large units, in class A only, the minimum number of primary samples should comply with the minimum number of units required for the laboratory sample (see Table 4).

Table 2. Number of randomly selected primary samples required for a given probability of detecting at least one non-compliance in a lot of meat or poultry product

Incidence of violative residues in the lot	Minimum number of samples (n0) required to detect a violative residue with a probability of:
%	90%	95%	99%
90	1	-	2
80	-	2	3
70	2	3	4
60	3	4	5
50	4	5	7
40	5	6	9
35	6	7	11
30	7	9	13
25	9	11	17
20	11	14	21
15	15	19	29
10	22	29	44
5	45	59	90
1	231	299	459
0.5	460	598	919
0.1	2302	2995	4603

Notes.

(a) The Table assumes random sampling.

(b) Where number of primary samples indicated in Table 2 is more than about 10% of units in the total lot, the number of primary samples taken may be fewer and should be calculated as follows:

where

n = minimum number of primary samples to be taken
n₀ = number of primary samples given in Table 2
N = number units, capable of yielding a primary sample, in the lot.

(c) Where a single primary sample is taken, the probability of detecting a violation is similar to the incidence of violative residues.

(d) This Table should not be used to determine the probability of detecting a violation in a lot of a plant product. As composite samples are prepared for plant products, the statistical distribution of residues in the lot must be known, to determine the probability.

Table 3. Meat and poultry products: Description of primary samples and minimum size of laboratory samples

	Commodity classification	Examples	Nature of primary sample to be taken	Minimum size of each laboratory sample
Class B, primary food commodities of animal origin
1.	Mammalian meats, type 06, group 030 Note: for enforcement of MRLs for fat soluble pesticides samples must be taken according to section 2 below.
1.1	Large mammals, whole or half carcass, usually 10 kg or more	cattle sheep pigs	whole or part of diaphragm, supplemented by cervical muscle, if necessary	0.5 kg
1.2	Small mammals whole carcass	rabbits	whole carcass or hind quarters	0.5 kg, after removal of skin and bone
1.3	Mammal meat parts, loose fresh/chilled/frozen packaged or otherwise	quarters chops steaks shoulders	whole unit(s), or a portion of a large unit	0.5 kg, after removal of bone
1.4	Mammal meat parts, bulk frozen	quarters chops	either a frozen cross-section of a container or the whole (or portions) of individual meat parts	0.5 kg, after removal of bone
2.	Mammalian fats, including carcass fat, type 06, group 031 Note: samples of fat taken as described in 2.1, 2.2 and 2.3 may be used to determine compliance of the fat or the whole product, with the corresponding MRLs
2.1	Large mammals, at slaughter, whole or half carcass usually 10 kg or more	cattle sheep pigs	kidney, abdominal or subcutaneous fat cut from one animal	0.5 kg
2.2	Small mammals, at slaughter, whole or half carcass <10 kg		abdominal or subcutaneous fat from one or more animals	0.5 kg
2.3	Mammal meat parts	legs chops steaks	either visible fat, trimmed from unit(s)	0.5 kg
			or whole unit(s) or portions of whole unit(s), where fat is not trimmable	2 kg
2.4	Mammal bulk fat tissue	-	units taken with a sampling device from at least 3 positions	0.5 kg
Class B, primary food commodities of animal origin
3.	Mammalian offals, type 06, group 032
3.1	Mammal liver, fresh/chilled/frozen	-	whole liver(s), or part of liver	0.4 kg
3.2	Mammal kidney, fresh/chilled/frozen	-	1 or both kidneys from 1 or more animal	0.2 kg
3.3	Mammal heart, fresh/chilled/frozen	-	Whole heart(s), or ventricle portion only, if large	0.4 kg
3.4	Other mammal offal, fresh/chilled/frozen	intestines brains	Part or whole unit from 1 or more animals, or a cross- section taken from bulk frozen product	0.5 kg
4.	Poultry meats, type 07, group 036 Note: for enforcement of MRLs for fat soluble pesticides samples must be taken according to section 5 below.
4.1	Bird, large-sized carcass >2 kg	turkey goose mature chicken	thighs, legs and other dark meat	0.5 kg after removal of skin and bone
4.2	Birds, medium-sized carcass 500 g-2 kg	duckling guinea fowl young chicken	thighs, legs or other dark meat from at least 3 birds	0.5 kg after removal of skin and bone
4.3	Birds, small-sized carcass <500 g carcass	quail pigeon	carcasses from at least 6 birds	0.2 kg of muscle tissue
4.4	Bird parts fresh/chilled/frozen, retail or wholesale packaged	legs quarters	packaged units, or individual parts	0.5 kg (after removal of skin and bone)
Class B, primary food commodities of animal origin
5.	Poultry fats, including carcass fat, type 07, group 037 Note: samples of fat taken as described in 5.1 and 5.2 may be used to determine compliance of the fat or the whole product, with the corresponding MRLs
5.1	Birds, at slaughter, whole or part-carcass	chickens turkeys	units of abdominal fat from at least 3 birds	0.5 kg
5.2	Bird meat parts	legs breast muscle	either visible fat, trimmed from unit(s)	0.5 kg
			or whole unit(s) or portions of whole unit(s), where fat is not trimmable	2 kg
5.3	Bird fat tissue in bulk	-	units taken with a sampling device from at least 3 positions	0.5 kg
6.	Poultry offals, type 07, group 038
6.1	Edible bird offal, except goose and duck fat liver and similar high value products		units from at least 6 birds, or a cross-section from a container	0.2 kg
6.2	Goose and duck fat liver and similar high value products		unit from 1 birds or container	0.05 kg
Class E, processed foods of animal origin
7.	Secondary food commodities of animal origin, type 16, group 080 dried meats Derived edible products of animal origin, type 17, group 085 processed animal fats Manufactured food (single ingredient) of animal origin, type 18 Manufactured food (multi-ingredient) of animal origin, type 19
7.1	Mammal or bird, comminuted, cooked canned, dried, rendered, or otherwise processed products, including multi-ingredient products	ham sausage minced beef chicken paste	packaged units, or a representative cross-section from a container, or units (including juices, if any) taken with a sampling device	0.5 kg or 2 kg if fat content <5%

Commodities are classified according to the Codex Alimentarius⁵
Refer to Table 1 to determine the number of primary samples required.

Table 4. Plant products: Description of primary samples and minimum size of laboratory samples

	Commodity classification	Examples	Nature of primary samples to be taken	Minimum size of each laboratory sample
Class A, primary food commodities of plant origin
1.	All fresh fruits, type 1, groups 001-008 All fresh vegetables, type 2, groups 009-019, except group 015 (dry pulses)
1.1	small sized fresh products units generally < 25 g	berries peas olives	whole units, or packages, or units taken with a sampling device	1 kg
1.2	medium sized fresh products units generally 25-250 g	apples oranges	whole units,	1 kg (at least 10 units)
1.3	large sized fresh products units generally > 250 g	cabbages cucumbers grapes(bunches)	whole units	2 kg (at least 5 units)
2.	Pulses, type 2, group 015	soya beans		1 kg
	Cereal grains, type 3, group 020	rice, wheat		1 kg
	Tree nuts, type 4, group 022	except coconuts		1 kg
		coconuts		5 units
	Oilseeds, type 4, group 023	peanuts		500 g
	Seeds for beverages and sweets, type 4, group 024	coffee beans		500 g
3.	Herbs, type 5, group 027	fresh parsley	whole units	0.5 kg
		others, fresh		0.2 kg
	(for dried herbs see: Class D, type 12, in section 5 of this Table)
	Spices, type 5, group 028	dried	whole units or taken with a sampling device	0.1 kg
Class C, primary animal feed commodities
4.	Primary feed commodities of plant origin, type 11
4.1	Legume animal feeds, and other forages and fodders		whole units, or units taken with a sampling device	1 kg (at least 10 units)
4.2	Straw, hay and other dried products		units taken with a sampling device	0.5 kg (at least 10 units)
Class D, processed foods of plant origin
5.	Secondary food commodities of plant origin, type 12, dried fruits, vegetables, herbs, milled cereal products Derived products of plant origin, type 13, teas, vegetable oils, juices, by-products for animal feed and miscellaneous products Manufactured foods (single ingredient) of plant origin, type 14 Manufactured foods (multi-ingredient) of plant origin, type 15, including products with ingredients of animal origin where the ingredient(s) of plant origin predominate(s), and group 078, breads
5.1	Products of high unit value		packages or units taken with a sampling device	0.1 kg*
5.2	Solid products of low bulk density	hops tea	packaged units, or units taken with a sampling device	0.2 kg
5.3	Other solid products	bread flour apple pomace dried fruit	packages or other whole units, or units taken with a sampling device	0.5 kg
5.4	Liquid products	vegetable oils juices	packaged units, or units taken with a sampling device	0.5 l or 0.5 kg
* A smaller laboratory sample may be taken from a product of exceptionally high value but the reason for doing so should be noted in the sampling record.

Commodities are classified according to the Codex Alimentarius⁵
Refer to Table 1 to determine the number of primary samples required.

Table 5. Egg and dairy products: Description of primary samples and minimum size of laboratory samples

	Commodity classification	Examples	Nature of primary samples to be taken	Minimum size of each laboratory sample
Class B, primary food commodities of animal origin
1.Poultry eggs, type 7, group 039
1.1	Eggs, except quail and similar, whole or otherwise		whole eggs, or units taken with a sampling device	12 whole chicken eggs, 6 whole goose or duck eggs
1.2	Eggs, quail and similar		whole eggs	24 whole eggs
Class E, processed foods of animal origin
2.	Secondary food commodities of animal origin, type 16, group 082 skimmed milks, evaporated milks and milk powders Derived edible products of animal origin, type 17, group 086 milkfats, group 087 butters, butteroils, creams, cream powders, caseins, etc. Manufactured food (single ingredient) of animal origin, type 18, group 090 Manufactured food (multi-ingredient) of animal origin, type 19, group 092 (including products with ingredients of plant origin where the ingredient(s) of animal origin predominates(s))
2.1	Liquid milks, milk powders, evaporated milks and creams, creams, dairy ice creams, yoghurts		packaged units, or units taken with a sampling device	0.5 l (liquid) or 0.5 kg (solid)
Notes.	(i) Evaporated milks and evaporated creams in bulk must be mixed thoroughly before sampling, scraping adhering material from the sides and bottom of containers and stirring well. About 2-3 l should be removed and again stirred well before removing the laboratory sample. (ii) Milk powders in bulk should be sampled by passing a dry borer tube through the powder at an even rate. (iii) Creams in bulk should be mixed thoroughly with a plunger before sampling but foaming, whipping and churning must be avoided.
2.2	Butter and butteroils	butter, whey butter, low fat spreads containing butter fat, anhydrous butteroil, anhydrous milkfat	whole or parts of packaged units, or units taken with a sampling device	0.2 kg or 0.2 l
Note.	Butter in bulk should be sampled with a minimum of 2 cores. Pats or rolls >250g should be quartered and opposite quarters taken as units.
2.3	Cheeses, including processed cheeses
	units 0.3 kg or greater		whole units, or units cut with a sampling device	0.5 kg
	units < 0.3 kg		whole units, or units cut with a sampling device	0.3 kg
Note.	Cheeses with a circular base should be sampled by making two cuts radiating from the centre. Cheeses with a rectangular base should be sampled by making two cuts parallel to the sides.
2.4	Liquid, frozen or dried egg products		units taken aseptically with a sampling device	0.5 kg

Commodities are classified according to the Codex Alimentarius⁵
Refer to Table 1 to determine the number of primary samples required.

Annex I. DEFINITION OF TERMS

Analytical portion

A representative quantity of material removed from the analytical sample, of proper size for measurement of the residue concentration.

Note. A sampling device may be used to withdraw the analytical portion.

The material prepared for analysis from the laboratory sample, by separation of the portion of the product to be analysed^5,6 and then by mixing, grinding, fine chopping, etc., for the removal of analytical portions with minimal sampling error.

Note. Preparation of the analytical sample must reflect the procedure used in setting Codex MRLs and thus the portion of the product to be analysed may include parts that are not normally consumed.

For plant products, the combined and well mixed aggregate of the primary samples taken from a lot. For meat, dairy and poultry products, the well mixed primary sample.

Notes.

(a) The primary samples must contribute sufficient material to enable all laboratory samples to be withdrawn from the bulk sample.

(b) Where separate laboratory samples are prepared during collection of the primary sample(s), the bulk sample is the conceptual sum of the laboratory samples, at the time of taking the samples from the lot.

The sample sent to, or received by, the laboratory. A representative quantity of material removed from the bulk sample.

Notes.

(a) The laboratory sample may be the whole or a part of the bulk sample.

(b) Units should not be cut or broken to produce the laboratory sample(s), except where subdivision of units is specified in Table 3.

(c) Replicate laboratory samples may be prepared.

A quantity of a food material delivered at one time and known, or presumed, by the sampling officer to have uniform characteristics such as origin, producer, variety, packer, type of packing, markings, consignor, etc. A suspect lot is one which, for any reason, is suspected to contain an excessive residue. A non-suspect lot is one for which there is no reason to suspect that it may contain an excessive residue.

Notes.

(a) Where a consignment is comprised of lots which can be identified as originating from different growers, etc., each lot should be considered separately.

(b) A consignment may consist of one or more lots.

(c) Where the size or boundary of each lot in a large consignment is not readily established, each one of a series of wagons, lorries, ship’s bays, etc., may be considered to be a separate lot.

(d) A lot may be mixed by grading or manufacturing processes, for example.

One or more units taken from one position in a lot.

Notes.

(a) The position from which a primary sample is taken in the lot should preferably be chosen randomly but, where this is physically impractical, it should be a random position in the accessible parts of the lot.

(b) The number of units required for a primary sample should be determined by the number of primary samples to be taken from the lot and by the minimum size and number of laboratory samples required.

(c) For plant, egg and dairy products, where more than one primary sample is taken from a lot, each should contribute an approximately similar proportion to the bulk sample.

(d) Units may be allocated randomly to replicate laboratory samples at the time of collecting the primary sample(s), in cases where the units are of medium or large size and mixing the bulk sample would not make the laboratory sample(s) more representative, or where the units (e.g. eggs, soft fruit) could be damaged by mixing.

(e) Where primary samples are taken at intervals during loading or unloading of a lot, the sampling “position” is a point in time.

(f) Units should not be cut or broken to produce the primary sample(s), except where subdivision of units is specified in Table 3.

One or more units selected from a population of units, or a portion of material selected from a larger quantity of material.

Sampling

The procedure used to draw and constitute a sample.

Sampling device

(i) A tool such as a scoop, dipper, borer, knife or spear, used to remove a unit from bulk material, from packages (such as drums, large cheeses) or from units of meat or poultry products which are too large to be taken as primary samples. (ii) A tool such as a riffle box, used to prepare a laboratory sample from a bulk sample, or to prepare an analytical portion from an analytical sample.

Notes.

(a) Specific sampling devices are described by ISO^1,2,3 and IDF⁴ standards.

(b) For materials such as loose straw or leaves, the hand of the sampling officer may be considered to be a sampling device.

A person trained in sampling procedures and, where required, authorised by the appropriate authorities to take samples.

Note. The sampling officer is responsible for all procedures leading to and including preparation, packing and shipping of the laboratory sample(s). The officer must understand that consistent adherence to the specified sampling procedures is necessary, must provide complete documentation for samples, and should collaborate closely with the laboratory.

The number of units, or quantity of material, constituting the sample.

Unit

The smallest discrete portion in a lot, which should be withdrawn to form the whole or part of a primary sample.

Note. Units should be identified as follows.

(a) Fresh fruit and vegetables. Each whole fruit, vegetable or natural bunch of them (e.g. grapes) should form a unit, except where these are small. Units of packaged small products may be identified as in (d), below. Where a sampling device may be used without damaging the material, units may be created by this means. Individual fresh fruit or vegetables must not be cut or broken to produce units.

(b) Large animals or parts or organs of them. A portion, or the whole, of a specified part or organ should form a unit. Parts or organs may be cut to form units.

(c) Small animals or parts or organs of them. Each whole animal or complete animal part or organ present may form a unit. Where packaged, units may be identified as in (d), below. Where a sampling device may be used without affecting residues, units may be created by this means.

(d) Packaged materials. The smallest discrete packages should be taken as units. Where the smallest packages are very large, they should be sampled as bulk, as in (e), below. Where the smallest packages are very small, a pack of packages may form the unit.

(e) Bulk materials and large packages (such as drums, cheeses, etc.) which are individually too large to be taken as primary samples. The units are created with a sampling device.

Annex II. SCHEMATIC REPRESENTATION OF SAMPLING

REFERENCES

1. International Organisation for Standardization, 1979. International Standard ISO 950: Cereals - Sampling (as grain).

2. International Organisation for Standardization, 1979. International Standard ISO 951: Pulses in bags - Sampling.

3. International Organisation for Standardization, 1980. International Standard ISO 1839: Sampling - Tea.

4. International Dairy Federation, 1985. International IDF Standard 50B: Milk and milk products - methods of sampling.

5. Joint FAO/WHO Food Standards Programme (1993). “Portion of commodities to which Codex Maximum Residue Limits apply and which is analysed”. Codex Alimentarius, Volume 2, Section 4.1, 389-404. FAO Rome. ISBN: 92-5-103271-8.

6. Joint FAO/WHO Food Standards Programme (1993). “Codex classification of foods and animal feeds”. Codex Alimentarius, Volume 2, Section 2, 147-366. FAO Rome. ISBN: 92-5-103271-8.