The Scottish Government

« Previous | Contents | Next »

Listen

MORTGAGE ARREARS AND REPOSSESSIONS IN SCOTLAND

Chapter Four: court data investigation

Objectives

4.1 The objectives of this phase of the research were:

• to collate and analyse 1994 and 2000 sheriff court records of repossession cases, and
• to collect contact details of borrowers who had had repossession actions raised against them. This was to contribute to the sample for the survey of borrowers.

4.2 Court registers and processes were investigated in four courts across Scotland; Glasgow, Edinburgh, Aberdeen and Falkirk. Two years were investigated, 1994 and 2000. One hundred per cent of court registers were consulted, and 16% of the relevant cases identified were randomly sampled from the court processes. The number of processes sampled in each court was roughly proportional to the number of repossession records in each of the courts in each of the years. However, as defended cases were of especial interest and these were in a minority, an additional 1% sample of these was drawn on top of the 16% random sample.

4.3. The research findings in this chapter refer to the nature of the case during the court action. It is not recorded if cases began with calling-up notices or letters of default, and how long this added to the formal repossession process

Choice of courts

4.4. Edinburgh and Glasgow were chosen as the biggest courts in Scotland to provide large numbers of records and to represent a significant proportion of Scotland's population. Aberdeen and Falkirk were chosen as they had large percentage increases between 1994 and 2000. It could be said that there is a possible bias towards cities; however, rural courts would give insufficient cases from which to generalise about Scotland as a whole unless an impracticable number were sampled. The two years were selected in order to collect information before and after the substantial rise that took place in the number of repossession decrees (Table 4.1).

4.5. The table shows the increase in the number of cases for each sheriff court. The percentage increases range from 6% to 2667% with a weighted mean of 209%. The very high level of overall increase and the remarkable extent of variation between the courts both appear to be questionable and indicate the need for this research. The small original numbers are obviously a factor in determining the largest increases by court.

Table 4.1 Repossession decrees granted in each court district 1994 and 2000

Source: Written answers to parliamentary questions by Mrs Margaret Ewing (Moray) (SNP) and Linda Fabiani (Central Scotland) (SNP) http://www.scottish.parliament.uk/official_report/wa-00/wa0629.htm

Table 4.2 The number of ordinary causes initiated and disposed of by nature of action, and whether the action was defended, during 1999

Notes

1. Source: Table 3.7, Civil Judicial Statistics 1999, Scottish Executive

2. CJS defines cases classified under mortgage lender as 'Repossession of property in breach/default of a house purchase loan. It is possible that relevant cases have been included under 'Land/Heritable'.

3. Defended actions include 'For pursuer Defended', 'Dismissal', 'Absolvitor', 'Remits'.

4.6. Table 4.2 shows that 7,790 cases classified under 'mortgage lender' were initiated during 1999. Seventy-six per cent (5,952) that were initiated reached a conclusion. Ninety-six per cent of cases (5,712) which reached a conclusion were granted in favour of the pursuer. Twenty-four per cent (1838) cases were abandoned before they reached a conclusion. This could be because the borrower paid, or rescheduled, their debt, or because vacant possession was yielded without the need for an order. Overall, the figures show that pursuers were very likely to be successful in obtaining orders. The number of initiated cases has not been published for each individual court.

Information in Court Records

4.7. There are two types of sheriff court record providing information on repossession cases: court registers, which summarise cases and court processes, which have the information in full.

4.8. Court registers contain summary information on individual civil law cases. While the registers are computerised at Edinburgh Sheriff Court for 1994 and 2000, and at Falkirk Sheriff Court (2000 only), they are held manually elsewhere.

4.9. Cases that might involve repossession were easy to identify in all the courts. Relevant cases were identified in the court registers from what was recorded in the 'crave'. Every action raised under the Conveyancing and Feudal Reform Scotland Act (1970) is required to cite the act in the crave. Therefore, one of two things was recorded under the craves; either the Conveyancing and Feudal Reform Scotland Act (1970), or 'recovery of possession of heritable property', or words to that effect. However, if an ejection order was pursued under the Section 5 Heritable Securities (1894) Act, as a follow-up to a calling-up notice being granted, then there could be a danger of failing to identify these cases as repossession cases because the crave may be unclear. Few (if any) repossession orders were likely to have been granted under the Heritable Securities (1894) Act as these only relate to mortgages granted before 1970 ⁸. All repossession cases were ordinary cause actions.

4.10 Court processes contain more detailed information on the characteristics and nature of the proceedings in each case. The information is recorded on paper. Processes are time consuming to scrutinise, given the amount of detail they contain, and the ease of access and use varies by court. Both Falkirk and Edinburgh Sheriff Courts allowed direct access to the record rooms, while initially in Aberdeen and Glasgow Sheriff Court a sheriff clerk representative acted as a gate-keeper and provided access on a case by case basis. The search was hindered by the fact that not all the processes were present in the archives in all the courts. Obviously, it was not possible to consult the processes that were not there and so a fresh sample was drawn.

4.11 A description of the information collected is included in Annex 1. This also contains a glossary of the legal terminology used in this report.

data collection

4.12 Permission was received to go into the courts in mid-February. An initial review of data in Edinburgh Sheriff Court was conducted in order to develop an understanding of the data and the way in which it could be structured and recorded for future analysis. An ACCESS database was then built and piloted. The data collection was undertaken in February - May 2001 by law students.

Collated Results

4.13 The total number of valid court records collated was 3,663: 487 in Aberdeen; 1049 in Edinburgh; 280 in Falkirk and 1,847 in Glasgow. All repossession cases were recorded from the registers, including those where no decree was granted.

4.14 In all four courts, there were fewer records in 1994 than 2000 as the following table shows.

Table 4.3 All residential repossession court cases initiated (research figures)

Notes

1. Commercial cases and cases with individuals, local authorities or Scottish Homes as the pursuer were removed.

2. This table includes all cases in which a decree was granted, as well as those where no decree was granted.

Discrepancies between Figures

4.15 There were discrepancies between the figures published in the CJS and those collated in this research.

4.16 Table 4.4 and Table 4.5 show the CJS figures supplied by the Scottish Executive compared with the figures collated in the primary data collection.

Table 4.4 CJS and research figures for repossession decrees initiated

Notes

Detailed monthly figures were available only for Glasgow 1994 enabling a detailed check to be carried out on this discrepancy. In this case, the difference is explained by the removal of cases from the research figures where local authority and individuals were pursuers. However, this would not explain the differences in the other courts and years, as the primary research mostly found more cases.

Table 4.5 CJS and research figures for repossession decrees granted

Notes

1. The 1994 CJS figures are from the Scottish Executive. Figures for the year 2000 are still being gathered and are not ready for publication. However, figures were compiled in order to answer Linda Fabiani's Parliamentary Question SIW-12907 & SIW-13498 and it is these that are presented in the table.

2. In both CJS and the research figures, only cases of granted decrees are shown.

3. These figures are not truly comparable as CJS note cases which end in 1994, and research figures note the end date of cases that began in 1994.

4.17 The differences in Table 4.5 do not follow a regular pattern. There were fewer records than expected in Glasgow and Falkirk in the year 2000. However, for 2000 for Aberdeen and Edinburgh and for the whole of 1994 in all four courts, more records were found than the CJS statistics suggested. The 15 cases for Aberdeen in 1994 contained in the CJS was surprisingly low. If the research figure is used instead, then the percentage increase between 1994 and 2000 shrinks from the original 1267% to 130%. This rise is still substantial but not on the scale the CJS figures suggest.

4.18 We have not been able to determine exactly why the research and the CJS figures vary so widely, although some factors are apparent that would, at least, introduce elements of approximation into the comparison. However, the effects of these factors could be to widen, as well as reduce, the discrepancies. The four principal factors we have found are detailed next.

Case classification errors

4.19 The CJS figures are collated from returns submitted by the courts under the Civil Judicial Statistics Act. When the courts make CJS returns, they do so from a set of codes applied to the actions at the registration stage. The relevant code which should be applied to cases of repossession is code H 'Mortgage Lender' - repossession of property in breach/default house purchase loan. However, there have been instances where codes were not correctly identified by court employees, and were entered under other codes by mistake. In particular, some cases were coded under code D 'Land or Heritable Estate' - recovery of debt out of heritable estate.

4.20 It should be noted that the CJS figures show that no cases were initiated in Aberdeen in 1994, but that 15 cases were granted in Aberdeen in 1994. The only explanation is staff error. It may be that the initiated cases were entered by the staff under a different heading from those that were recorded as granted under mortgage lender.

Difficulty in distinguishing between commercial and residential repossessions

4.21 Legal advice and advice from court officers suggested that it is often impossible to identify the difference between a residential and a commercial repossession case. The only difference may be that in a commercial context the case is more likely to begin with a Calling-up Notice. When possible, commercial cases were identified and removed when inspecting and editing our database of court records. For instance, entries were removed when the pursuer was a brewery and the property addresses was of a public house. For some commercial cases, it would not be possible to tell from the address, and so these will remain in the database ⁹.

4.22 Despite these edits, a few repossession decrees for commercial properties are probably still included in our figures. Our expectation is that the CJS records will not have edited out the commercial cases, because there was no requirement for them to do so. This would have the effect of widening the discrepancies between the CJS and the research figures.

Removal of cases with individuals, local authorities or Scottish Homes as the pursuer

4.23 Cases where the pursuer was a local authority or Scottish Homes were also removed. It is highly unlikely that these cases were to do with rent arrears that were misplaced. Rent arrears are not included under the Conveyancing and Feudal Reform Scotland Act, (1970). Rent arrears cases are almost always raised under summary cause procedure, although very occasionally under ordinary cause procedure, usually where the amount of arrears is more than 1500. If the case were to do with rent arrears, this would normally be cited in the crave. Local Authorities would not normally use the Conveyancing and Feudal Reform Scotland Act, (1970) to evict a tenant, as they could use the Housing (Scotland) Act 1987, which would give them the power to evict.

4.24 It could be that these cases were either local authority mortgages or discounted RTB sales. In a discounted RTB sale, the local authority would hold a standard security over the discounted part of the sale. The proportion of cases removed from the court records with local authority pursuers was less than 5%. However, the proportion of loans with local authority mortgages is not known.

4.25 However, it was assumed that the cases under the Conveyancing and Feudal Reform Scotland Act, (1970) where a local authority was the pursuer would concern a commercial property, and these were removed.

4.26 Cases where the pursuer was an individual were also removed from the analysis. Removing these from the database must have the effect of increasing the discrepancy between the CJS and the research figures.

Year differences

4.27 The comparisons between the CJS and the research data are of granted cases. These are recorded by end date by the courts, i.e. the cases reported for a year will have concluded in that year. However, the research had to collate all cases, not only those that were granted. End year dates are only set when the case is granted so the research had to sample and report cases that were initiated in the year. This will have some effect on comparability, but as cases are usually over within one or two months this should not be significant. Such effect as there is will be to narrow the gap between the research and CJS figures, because the research data will be of a slightly later period at a time of rising numbers.

4.28 It has not been possible to resolve the differences between the repossession figures in a conclusive way. It seems likely, though no firm evidence exists to support the hypothesis, that recording court data has not always been carried out consistently across Scotland. Two factors suggest this: firstly, the data presented by court in Table 4.1 raises obvious questions of consistency. It seems most improbable that the variations by year and court in the volume of decrees granted reflect an authentic pattern. Secondly, the comparisons with the research data shown in Table 4.4 and Table 4.5 are also very irregular. This implies that the differences are not procedurally based or systematic in any other way. It can be asked whether the research data is a sound benchmark. While we believe this is so, it cannot be conclusively demonstrated: however, it is certain that the procedures and definitions were consistent across all the courts so we feel confident in attributing most, or all, of the irregularity in the comparisons to the CJS data.

4.29 This makes it difficult to draw firm conclusions about the total numbers of repossession cases granted by courts. Our best estimate for the courts we surveyed must be that the numbers are higher than in the CJS returns, but there is no sound basis for quantifying this across Scotland. Inconsistencies in the data relationships indicate that the findings from the sampled courts are unlikely to be more widely applicable.

sampling of processes

4.30 A problem in all the courts was that some of the processes were not present in the archives. It was suggested at the courts that some of these files could still be with solicitors. However, the information in the court registers suggested that the missing processes were not distinctive in any way.

4.31 Table 4.6 shows the number of processes examined in each year in each court.

Table 4.6 Numbers of processes sampled

Analysis of the court data

Decrees

4.32 In the main, decrees are distinguished between those 'for pursuer (undefended)'; 'for pursuer (defended)'; 'dismissal'; 'absolvitor' and 'sist'. A glossary of terms is contained in Annex 1. By far the most common type of decree was for the pursuer (i.e. the bank or building society). Generally, the cases were undefended, as shown in Table 4.7. There were also 135 defended cases where the decrees were granted for the pursuers. Defended cases were identified by the case outcome, whether a 'Notice of Intention to defend' (NID) was lodged or whether a defender solicitor was recorded. The NID indicates that the defender intends to defend the action, but does not contain any details of the proposed defence. The defender must then lodge defences with the court within 14 days of the expiry of the period of notice of the action (which is usually 21 days after service). Where defences are not lodged, or the defender fails to turn up or be represented, decree by default is granted.

4.33 However, it became evident from the study of court processes that typically, even when a 'Notice of Intention to defend' (NID) was lodged with the court, no defences were lodged, or less commonly, the defender failed to show up at the hearing. Therefore, although the cases were recorded as defended, no defence was actually lodged in many cases. The data is given for the 2 years combined rather than by year in Table 4.7.

Table 4.7 Repossession case outcomes (1994 and 2000)

Notes

1. 747 cases were abandoned. 'No date given' was entered into the database when the case had no end date recorded in the registers. This normally indicated that no decree was granted and the case was abandoned, but it is possible that the date was not entered into the registers by the court staff, either because it had been forgotten or because the case had not yet completed. This could only be checked in those cases where processes were sampled. For the cases where the processes were checked, no date indicated either a settlement out of court, or the case being abandoned.

2. A glossary of legal terminology is contained in Annex 1.

3. There was a great variety of case outcomes, some of which had only one case in a category. These have all been subsumed within the categories shown in the table above.

4. 'By default' is recorded when defences are not lodged, or the defender fails to turn up or be represented.

4.34 The case outcomes have been combined as granted in favour of the pursuers and not so granted. Table 4.8 shows that in over three-quarters of cases a decree was granted in favour of the pursuer. However, most of the other cases were not resolved in favour of the defender. It is possible that fewer decrees in favour of the pursuer will be granted under the new act - Mortgage Rights (Scotland) Act 2001.

Table 4.8 Outcomes of cases (1994 and 2000)

Notes

* Cases which were granted in favour of the Pursuer, includes those which were defended and not defended, as well as cases granted 'By Default'.

# 'Not granted' includes dismissals, absolvitor, cases where no date is given in the court registers, any sist cases, and cases where they were settled out of court or delayed. The recording of case outcomes was far from consistent in the court registers and the great variety of recorded case outcomes was summarised.

4.35 Further results are presented in Tables 4.9 and 4.10. Decrees were less likely to be granted in favour of the pursuer in cases which were defended. The figures show that a higher proportion of cases were granted in favour of the pursuer in 2000 than in 1994 (Table 4.9).

Table 4.9 Start year by outcome of repossession court case

Note

Significant at the 99% level

4.36 Slightly more cases were granted in favour of the pursuer in Falkirk than in the other courts. Cases were least likely to be granted in Glasgow Sheriff court.

Table 4.10 Location of cases (1994 and 2000) by outcome

Note

Significant at the 95% level

Pursuers

4.37 The most common pursuer to appear in the search of the records was the Bank of Scotland (now HBoS). Together, Bank of Scotland (now HBoS), Abbey National plc, Halifax plc (now HBoS), Clydesdale Bank plc, Lloyds TSB Scotland plc and Alliance & Leicester plc accounted for over 50% of all the cases of repossession examined in the four Courts (Table 4.11).

Table 4.11 Pursuers of repossession cases (1994 and 2000)

Notes: TSB Scotland and Lloyds have been changed into one category 'Lloyds TSB Scotland plc'. Lenders with a count of less than 100 have been excluded from this table.

4.38 The market share of each lender was obtained from the Land Register for all lenders with standard securities for properties sold in 2000. A comparison of market share in 2000 and court decrees granted in 2000 was considered, but it was determined that the data were insufficiently comparable. The market share of lenders is shown in Table 5.2.

4.39 Land Register data reflected market shares in 2000, which was an indication of the most recent flow of loans, whereas the repossession actions relate to the market share over an earlier period. Changes in the market environment - products, marketing strategy, and mergers - may therefore invalidate comparisons. Furthermore, the court data were collected from four courts and do not represent the whole of Scotland. Comparisons may be unduly influenced by differences in the presence of these institutions in particular geographic areas.

Use of Solicitors

4.40 To formally initiate court action, the pursuer needs legal representation. This is not the case for the defender, and few defenders (3%) had solicitors. The information on solicitors illustrates that just under 65% of repossession cases we came across were handled by 13 solicitor firms in Scotland and occasionally in-house by lenders. This is because lenders tend to use the same firms of solicitors and solicitors tend to act on behalf of a number of lenders.

4.41 Even if a solicitor was not identified, it could still be recorded that a defence took place.

4.42 It should not be inferred that providing borrowers with legal representation would materially alter case outcomes. As the law stood before the Mortgage Rights (Scotland) Act 2001 there was little chance of a borrower's circumstances being taken into consideration in court. The only 'defence' possible was the defender denying that the money was owed. Therefore, in most cases, there was little point in a defender employing a solicitor. There was little chance of being awarded legal aid because of the likelihood of losing the case.

4.43 Very few cases (6%) were recorded as defended. The figures include cases where a NID was lodged and cases where a defence solicitor or party defending ¹⁰ was entered. Even this low recorded level certainly exaggerates the number of cases where a defender or defender's representative was actually present.

4.44 There were appreciable differences (significant at 99% level) between courts in the proportion of cases that were defended, with the highest proportion being in Falkirk (Table 4.12). It is not known if there are any specific reasons, particular to Falkirk court, that could account for this difference.

Table 4.12 Percentage of defended cases by location of the court

Note

Defended cases were identified by the case outcome, whether a 'Notice of Intention to defend' (NID) was lodged or whether a defender solicitor was recorded.

4.45 There was a slight difference between the outcomes of undefended and defended cases. 'Defended' cases were less likely to be granted (see Table 4.13).

Table 4.13 Outcome of cases by whether defended or not defended

Notes

1. Significant at 95% level

2. Defended cases were identified by the case outcome, whether a 'Notice of Intention to defend' (NID) was lodged or whether a defender solicitor was recorded.

4.46 Since defended cases are of particular interest, a list of all the defended cases with the locations and the outcomes is presented in Annex 2.

Legal Aid

4.47 None of the defenders obtained legal aid in any of the processes examined. Legal aid was unlikely to be granted in a case of repossession as the law (before the introduction of the Mortgage Rights (Scotland) Act 2001), allows little protection for the borrower in arrears. The Scottish Legal Aid Board has provided figures for legal aid applications for repossession court cases and these are shown in the table below. Despite indications to the contrary, Table 4.14 shows that a small number of people were granted legal aid to pursue repossession court cases. It was also clear that many more cases were refused legal aid than granted. Most importantly, the information from the Scottish Legal Aid Board has shown just how few people have even applied for legal aid in the recent past.

Table 4.14 Legal aid applications for repossession court cases

Notes

1. These figures were obtained from the Scottish Legal Aid Board.

2. Cases were identified as being for the recovery of heritable property, and cases where banks and building societies were the opponents were manually identified.

3. Cases that had councils or housing associations as opponents and cases where the opponent was an individual were excluded.

4.48 It remains to be seen how the advent of the Mortgage Rights (Scotland) Act 2001 will affect the above figures, which are presented here as a baseline measure.

Length of Cases

4.49 The date when the court proceedings began is always recorded in the court register and when a decree is granted, either for pursuer or a dismissal, an end date is also available in many cases. It was possible to calculate the length of the proceedings for 80% of the cases in the study. The start and end date refers only to the length of time the case takes in court. It is not possible to discover how long the entire process, as experienced by the parties involved, takes from initiation to conclusion. The part of the process that can be measured is from when the application for the Section 24 warrant is lodged with the court to when a decree is granted. This process does not include the time taken for the Calling up Notice or the Letter of default. Neither of these are registered with the court and it is not known how many are applied for and how long this adds to the repossession process. However, not all lenders use either of these methods, as some apply straight away for a section 24 warrant.

4.50 A repossession court case is nearly always short, with just under 90% of cases being completed within 3 months. While the causes of the difference in length of cases may not be immediately obvious, they may be partly explained by the levels of business in each particular court. It may also be partly related to the numbers of defended cases, which by their nature take longer. More detail on the length of cases and the variation between courts is contained in Annex 3.

Age of mortgage when court action began

4.51 Most of the court processes showed the date when the standard security was recorded in the Land Register. This date, together with the date when the court action is initiated, gives an indication of the length of loan before the borrower got into difficulty. Based on a sample of 579 cases, court action was initiated by the lender within 5 years of the house being purchased in nearly 70% of cases (see Table 4.15). In over a quarter of cases, this occurred within two years. It should be noted that a new standard security would be recorded in the Land Register if a re-mortgage took place, so it cannot be certain that this record always indicates the length of time the borrower has owned the house. Nevertheless, this information is consistent with that provided by the lenders.

Table 4.15 Length of mortgage (before court action began)

4.52 This picture of court action occurring early in the loan period is to be expected since it is known the burden of payments is highest at this time. Mortgages that start with a lower fixed rate of interest seem to have become more common, so the sudden switch to a higher rate after a year or two may cause problems for some. Also, it is usually true that people get pay rises and progress in their jobs as their mortgage progresses and, of course, inflation reduces the payment burden over time.

Trends and variations

4.53 Case outcomes, duration, and length of mortgage were analysed for each of the courts. None showed differences from the general picture, apart from the lengths of the court cases. The variation in case duration between courts is shown in Annex 3.

4.54 There was no significant difference between courts in the length of mortgage before court action began.

4.55 Trends between 1994 and 2000 showed an increase in the total number of repossession court cases between 1994 and 2000 (Table 4.16). This applied to cases initiated and cases granted. Overall, the proportion of cases granted has increased between 1994 and 2000. However, the proportion of cases granted varies between courts. Aberdeen and Glasgow have a higher proportion granted than in 1994, while Edinburgh and Falkirk show little difference. Table 4.17, based on the CJS figures, shows a more erratic pattern.

Table 4.16 Difference between 1994 and 2000 figures by court

Note

All these figures were taken from the primary data collection and not from the CJS.

Table 4.17 Difference between 1994 and 2000 figures by court (CJS figures)

implications

4.56 The implications of the findings presented in this chapter will be brought together in the synthesis presented in Chapter 8. The principal points to be taken forward there are:

• the CJS records by court and year (Table 4.1) appear erratic. They suggest inconsistencies in the way in which data have been defined and recorded between courts and over time
• the CJS figures also have an erratic relationship with the research figures collated at the sampled courts (Table 4.4 and Table 4.5). The research data were collated and recorded in a consistent way at all courts. This implies that the discrepancies between the figures do not have a systematic cause, but result from practices that are inconsistent between individual courts and years
• the Scotland-wide CJS indicator shows a higher rate of increase between 1994 and 2000 than the CML data. The high CJS trend could be partly a result of 1994 baseline records being erroneously low. Some of the 1994 figures for individual courts are surprisingly low and the only one of these sampled, Aberdeen, showed an increase from 15 CJS records to 92 collated in the research
• the 1994-2000 rate of increase was 139% for all courts, 100% for the sampled courts, 165% for the unsampled courts and 70% for the research data at the sampled courts. The rate of increase in the CML records of repossessions (rather than orders granted) is also 70%
• CML member's policies and practices are unlikely to vary geographically, which supports the assessment that the CJS discrepancy in trends between the sampled and unsampled courts is likely to reflect inconsistent data collation practices
• while the research data is based on a sample of courts, this is unlikely to undermine comparability with the CML results, where there is no evidence of significant geographic error or bias. The precise matching (70%) between the sources may be a little fortuitous but, nevertheless, it strongly supports the judgement that the CML data are more consistent and meaningful than the CJS data. It also suggests that the figure of 70% is close to a true record of the increase between 1994 and 1999
• there are differences between the research and CJS data recording that will tend to increase the discrepancies. The research excluded records where there was evidence that commercial properties were involved or that pursuers were individuals, Scottish Homes or local authorities
• very few cases were dismissed (only about 2%), supporting the view that borrowers in arrears had little chance of success in court, as the law then stood
• few cases were defended and no borrower had legal aid
• repossession cases from the time the case comes to court were short, usually being completed within two months. The length of time from the calling-up notice and letter of default is not recorded in official sources

repossession actions were most commonly raised when loans were relatively young. Most (76%) loans were less than 6 years old and only 7% were over 7 years old.

« Previous | Contents | Next »