Skip to content

Portugal · Interlibrary exchange · 2024

Net lenders can also be borrowers: Portugal’s interlibrary flows

Of 187 municipal library services reporting positive interlibrary flows, 146 reported both lending and receiving. Converting each pair into a net balance removes 48.00% of summed directional volume. That arithmetic does not establish reciprocal partners or unique network loans.

Published
Share
Share this page
EmailCite & reuse this study
Key findings as text
Reusable imagesNet lenders can also be borrowers: Portugal’s interlibrary flows — 146 / 187: Positive-flow services reporting both directions. Municipal RNBP records, not individual library buildings.Download sharing image (JPG) ↓

Key findings

146 / 187

Positive-flow services reporting both directions

Municipal RNBP records, not individual library buildings.

48.00%

Gross directional volume removed by local netting

Arithmetic cancellation, not matched or reciprocal transfers.

7 / 257

Municipal records missing both flow values

Excluded from paired calculations, never treated as zero.

50.00%

Median municipal netting cancellation share

187 positive-flow records; differs from gross-volume weighting.

A net lender may still need other libraries

A library can lend more documents to other libraries than it receives and still borrow a substantial amount itself. Calling it a net lender answers which direction is larger. It says surprisingly little about the smaller direction, the amount of activity, or the relationships that make the service work. A single balance compresses two useful numbers into one.

The 2024 statistics for Portugal's Rede Nacional de Bibliotecas Públicas, or RNBP, provide a concrete example. Among 187 municipal library services reporting positive interlibrary activity in at least one direction, 146 reported documents both lent and received. That is 78.1% of the positive-activity group. A description built around one-way supplier and recipient roles would miss the dominant pattern in these records.

This study pairs the two published municipal fields, keeps zero and missing responses separate, and asks what disappears when they are reduced to a net balance. It also checks whether the answer changes when large services receive more weight. The result is a picture of reported directional activity in one year, not a map of partnerships, an assessment of service quality, or a history of permanent lending roles.

What the records actually cover

The source is DGLAB's official RNBP dataset for 2024, supplied by the municipalities and released through Portugal's open-data portal. The file contains 257 municipal records. The reporting unit can encompass a municipal library network, including its branches and other facilities. It is therefore important to say municipal services rather than treat each row as one library building.

The questionnaire separates home lending in section 8 from interlibrary lending in section 9. The two fields used here are total documents lent and total documents received through interlibrary borrowing. They are annual reported counts. The data do not identify the documents, their subject matter, the readers using them, the institutions on the other end, or the requests that could not be fulfilled.

Both selected fields contain values for 250 municipalities. The same seven municipalities have neither value, so there are no records with only one direction missing. All 514 selected cells were checked across the source CSV and spreadsheet. The marginal counts and sums also reconcile with the official summary. That confirms this extraction, while leaving the usual limits of administrative self-reporting intact.

Zero activity and missing activity belong in different boxes

Of the 250 paired records, 63 report zero in both directions. The remaining 187 report at least one positive count. Among these, 36 report documents sent but none received, while five report documents received but none sent. Another 79 report both directions with more sent, and 57 report both with more received. Ten report the same positive count in each direction.

These categories are mutually exclusive and account for every paired record. Adding the seven missing pairs completes the original 257-record frame. Keeping that final group visible prevents a familiar mistake: converting a blank into evidence of no activity. The source does not tell us what the missing counts would have been, and there is no assumption here that the omissions are random.

Even an explicit zero has a limited meaning. It is a report of no documents in that category during the year. It does not establish that the service was unavailable, that no request was made, or that a library would refuse a future request. The categories describe recorded activity, not the full offer available to readers.

The two directions behind a net role

Subtracting received documents from sent documents produces 115 net-sending municipalities, 62 net-receiving municipalities and ten with a positive balanced flow. The 63 zero-activity records are separate: although their numerical balance is also zero, they are not examples of equally active lending and borrowing. Combining those two kinds of zero would blur the most basic distinction in the data.

Among the 115 net senders, 79 also received documents. Among the 62 net receivers, 57 also sent documents. Those proportions are 68.7% and 91.9%, respectively. The terms net sender and net receiver remain mathematically correct, but neither can safely be shortened to sender-only or receiver-only. The opposite direction is present in most cases in both groups.

There is a practical reason to retain both numbers in a public dashboard. A reader seeing a net balance of twenty cannot tell whether it comes from twenty sent and none received, or from 120 sent and 100 received. The balance is identical; the recorded activity is not. Neither example automatically describes a better service, but they plainly describe different patterns.

Measuring how much netting removes

For each municipality, gross directional volume is the sum of documents sent and received. Net balance is their difference. The absolute net balance keeps the size of that difference while ignoring which direction is larger. Subtracting the absolute net balance from gross volume gives the amount removed when the two directions cancel in the calculation.

Consider a deliberately simple example: eighty documents sent and twenty received. Gross volume is one hundred and net balance is sixty. Netting removes forty, or 40% of gross volume. Algebraically, the removed amount is twice the smaller directional count. This is an arithmetic decomposition. It does not match particular documents or demonstrate that the same partner sent something back.

The netting cancellation share runs from zero for a positive one-way record to 100% for equal positive counts. It is undefined when both counts are zero, because there is no positive volume to divide by. Those records remain in the coverage results but are excluded from this ratio. The share tells us how much reported directional volume netting removes, not how cooperative, efficient or successful a service is.

Counting municipalities and counting volume answer different questions

Across the positive-flow records, the reported sent and received totals are 13,197 and 9,040. Their sum is 22,237 directional counts. Within-municipality netting removes 10,674 of that gross total, giving a volume-weighted cancellation share of 48.00%. Nearly half the gross directional volume disappears when each municipality is represented only by the size of its net balance.

The median municipal cancellation share is 50%. This is a different summary: it gives each of the 187 positive-flow municipalities one position in the ordered distribution, regardless of how many documents it reports. The volume-weighted result gives a municipality more influence when its gross count is larger. Neither calculation should be substituted for the other without saying what is being weighted.

Nor should the gross sum be called 22,237 unique loans across a national network. If both participating services are in the frame, one interlibrary loan may contribute to a sent count at one endpoint and a received count at the other. Without partner and transaction identifiers, the data cannot remove that duplication or reconstruct distinct transfers. Gross volume here is deliberately a sum of reported directions.

An almost balanced total can hide uneven local balances

The 146 municipalities reporting both directions provide a useful second view. Together they report 9,352 documents sent and 8,951 received. Their aggregate difference is just 401, compared with a gross directional total of 18,303. Looking only at the aggregate difference might suggest a broadly balanced pattern throughout the group.

That interpretation does not follow. The median cancellation share within this two-way group is 64.52%, while its volume-weighted share is 58.32%. Individual positive and negative balances can offset when summed across municipalities. A small final group balance therefore does not mean that each service sends and receives similar numbers. Aggregating first and examining local differences first are different operations.

This distinction matters whenever administrative flows are summarized into one national or regional number. It is possible for contrasting local roles to coexist under an apparently balanced headline. The paired municipal data make that contrast visible, but they still do not show which municipalities exchanged documents with one another. A balanced group total cannot supply those missing links.

How strongly one-sided is one-sided?

A net balance divided by gross volume places every positive-flow record between minus one and plus one. The endpoints represent activity in only one direction, while zero represents equal positive counts. This scale provides a way to examine the strength of the imbalance without confusing a large absolute count with a large proportional difference.

Using an absolute normalized balance of at least 0.50 identifies 98 of the 187 positive-flow municipalities. Raising the threshold to 0.80 reduces the count to 57. These are explicitly chosen analytical cutoffs, not professional standards. A municipality does not become deficient merely by crossing one of them, and the thresholds do not describe an official classification.

Among the 146 two-way records alone, 57 meet the 0.50 cutoff and sixteen meet the 0.80 cutoff. Reporting both counts makes the sensitivity clear. A description such as strongly one-sided has little meaning unless the rule is supplied. Keeping the continuous balance alongside the categories also lets readers see that a tiny numerical change near a cutoff need not represent a meaningful service change.

A few large records influence the volume-weighted picture

The largest gross-flow municipality accounts for 12.17% of the positive-flow total. The largest five account for 30.98%, and the largest ten for 42.62%. This concentration explains why a volume-weighted statistic can be influenced by a relatively small number of records even though many municipalities participate in the survey.

As a predefined sensitivity check, the calculation is repeated once without the largest gross-flow record. The weighted cancellation share rises from 48.00% to 54.65%, while the median municipal share remains 50%. The difference is informative: the summary based on volume is more sensitive to this particular record than the middle of the municipal distribution.

The record is retained in every main result. There is no evidence here that it is erroneous, and being influential is not a reason to delete it. This check is also not an anonymous criticism of its library service. It shows how weighting works. A sound presentation can report the full result, identify the sensitivity and resist replacing the main analysis with whichever version makes a tidier story.

Why the national directions do not have to add up

The difference between the reported sent and received totals is 4,157. That number is not evidence of missing physical documents, failed returns or bookkeeping misconduct. The dataset covers a particular reporting frame rather than every possible lending partner. A municipal service can exchange material with institutions outside that frame, and the observed rows do not identify those counterparties.

The survey also contains seven missing paired records, while definitions and reporting boundaries can affect administrative totals. Those possibilities justify caution; they do not let this study attribute the whole imbalance to any single explanation. In particular, 4,157 cannot be relabelled as a verified number of documents sent outside the RNBP. The necessary partner-level evidence is absent.

It is equally important to distinguish receiving a document on loan from the return journey of a document previously lent. The questionnaire's two directions describe lending and borrowing categories, not a parcel-tracking system. Reading their difference as outstanding unreturned books would change the meaning of the measure and create an unsupported claim about library operations.

What a better public summary would show

A compact dashboard can preserve useful information without becoming a technical report. It can show the number of paired responses, the number missing, the number with zero reported activity, and the six directional categories. For services with positive counts, it can display sent and received values together before offering the convenience of a net label.

A second view can place gross volume beside normalized balance. This separates the amount of reported activity from its directional tilt. The figure accompanying this study retains the full range, including the largest record, instead of trimming it to make the smaller records easier to arrange. A logarithmic volume axis is labelled explicitly so that spacing is not mistaken for a linear difference.

For a quick numerical summary, the median municipal cancellation share and the gross-weighted share can appear side by side, each with its denominator. A short sensitivity note can explain the effect of the largest record. These additions help a reader understand what the chosen summary emphasizes without claiming that one statistic supplies a complete account of cooperation.

What would be needed to go further

The next step for a study of persistent lending roles would be harmonized municipal records over several years, stable service identifiers and a clear account of reorganizations or reporting changes. One year's net direction cannot establish that a municipality is usually a supplier or usually a recipient. A temporary project or an unusual year's requests could matter, but this dataset cannot test those explanations.

A study of actual reciprocity would need counterparties and matched loan records. A study of service performance would additionally need requests, fulfilment outcomes, delivery times and relevant resource information. Reader benefit would require evidence about use and experience. None of those questions can be settled by adding interpretive labels to the two annual flow columns used here.

The narrower conclusion is still useful. In this RNBP reporting frame, most municipalities with positive interlibrary activity report both directions, and almost half of the summed directional volume is removed by within-municipality netting. Net roles offer a convenient summary, provided the two underlying flows remain available. When they disappear, a simple label can make an active borrower look like only a lender.

Net roles hide two-way interlibrary activity

Portugal RNBP 2024 · 257 municipal records · Zero and missing remain separate

Full chart and source
Category counts: two-way net sent 79; two-way net received 57; positive balanced 10; sent only 36; received only 5; zero in both 63; both missing 7.

Scroll horizontally to inspect the chart. ↔

Chart and values

Scroll horizontally to see all columns.

Every municipal record, with zero and missing separate
Reported direction categoryMunicipalities
Two-way, more sent79
Two-way, more received57
Two-way, equal positive counts10
Sent only36
Received only5
Zero in both directions63
Both values missing7

257 municipal services; the first five rows cover 187 positive-flow records. Categories are mutually exclusive.

Mutually exclusive categories, including 63 zero records and seven missing pairs. A net sender may also receive documents.

Download chart (SVG) ↓ · Download chart (PNG) ↓

Direction and volume are different dimensions

All 187 positive-flow municipalities · Horizontal balance; logarithmic gross-volume axis

Full chart and source
Scatter plot of all 187 positive-flow municipal records, normalized balance from minus one to plus one and gross volume on a labelled log scale from one to 3,000. Full coordinates are in the accompanying table.

Scroll horizontally to inspect the chart. ↔

Chart and values

Scroll horizontally to see all columns.

Positive-flow records underlying the scatter plot
Municipal codeSentReceivedGrossNormalized balance
304019215881680−0.890476
20101129411700.517647
403015763120−0.050000
20102242268510−0.050980
50101187250.440000
40101202220.818182
3040264100.200000
30101129471760.465909
3010295531480.283784
10801145−0.600000
40401256691−0.450549
20801151833−0.090909
30501128131410.815603
404024539840.071429
305024566111−0.189189
40403355085−0.176471
105011011.000000
10201055−1.000000
201032091213300.266667
204021011.000000
101012416400.200000
2040397491460.328767
106013140.500000
4030214135149−0.812081
30103207397604−0.314570
20104181582390.514644
40203132205337−0.216617
305036617830.590361
102022816440.272727
3030351015−0.333333
40405198324522−0.241379
30504213152−0.192308
403031851853700.000000
108022626520.000000
102032707027071.000000
1080313001301.000000
3010575261010.485149
30106128187315−0.187302
20602124160.500000
30505123547−0.489362
30202372390.897436
20404318390.589744
20901044−1.000000
10504145−0.600000
501046061.000000
4040674401140.298246
208023938770.012987
30506404989−0.101124
105065051.000000
20202930931.000000
30403222240.833333
30507365288−0.181818
210024083123−0.349593
402065051.000000
404072893121−0.537190
40207110111.000000
304044041.000000
104011011.000000
20105173812540.362205
40304308380.578947
10301450214710.910828
50105254065−0.230769
10507103130.538462
40408213354−0.222222
3040519194213−0.821596
2020421012−0.666667
20405133160.625000
20701286340.647059
2100372681400.028571
204064150.600000
305081510250.200000
207023113440.409091
4010215731600.962500
208046211730.698630
10302175220.545455
20106246410656−0.250000
50106274443180.723270
5010773116189−0.227513
203024435651008−0.121032
501082421604020.203980
30204196250.520000
30107420554750.768421
2040736003601.000000
105082240.000000
205013140.500000
301083360.000000
104033031.000000
20604160161.000000
20805151934−0.117647
10404130131.000000
2010765110.090909
2080692332−0.437500
10103244102540.921260
40409167291960.704082
204082240.000000
108066061.000000
108082130.333333
303045051.000000
10104347−0.142857
108092240.000000
40306255681−0.382716
202063140.500000
303052021.000000
2060591625−0.280000
404105818760.526316
3010928265293−0.808874
2060641822−0.636364
30110157085−0.647059
401035813−0.230769
30205454490.836735
50110424486−0.023256
106034646920.000000
201092532044570.107221
30411268−0.500000
40411202545−0.111111
20110372467839−0.113230
303064041.000000
20410202220.818182
1051202020−1.000000
10105300301.000000
20411180181.000000
10513212230.826087
2060873100.400000
204122021.000000
30111305282−0.268293
208075233850.223529
2030488160.000000
1010691221−0.142857
10107650651.000000
40214122140.714286
40308598670.761194
50111102535−0.428571
104051510250.200000
20305112130.692308
10406100101.000000
403096814−0.142857
403104012520.538462
580084040800.000000
30509122381600.525000
30510203392420.677686
5800905050−1.000000
106024260.333333
3051113321350.970370
401048081.000000
1030412481320.878788
501125124750.360000
1060424502451.000000
10712120121.000000
2061001010−1.000000
580113030600.000000
304074610−0.200000
2070367601270.055118
30307150151.000000
107132312350.314286
4041295561510.258278
205044812−0.333333
303083031.000000
30309102120.666667
201118891179−0.016760
50113106160.250000
30112132671990.326633
403111011.000000
50114142135−0.200000
30408142160.750000
20612270271.000000
304097071.000000
301133311454760.390756
20809119200.100000
20112121461670.449102
10108122739−0.384615
104073031.000000
4031221719−0.789474
4031342135177−0.525424
1010982100.600000
40413220221.000000
103052130.333333
1040864100.200000
107173031.000000
302071011.000000
10110107170.176471
10306235−0.200000
1040921517−0.764706
10719134−0.500000
50116125769−0.652174
102064260.333333
1081351217−0.411765
206142915440.318182

All 187 positive-flow records, identified by source municipal code. Counts are reported directions, not unique network transfers. Ratios rounded only for display.

Logarithmic volume axis, no clipped tail. Point overlap is possible. Zero/zero and missing records are excluded only because normalized balance is undefined.

Download chart (SVG) ↓ · Download chart (PNG) ↓

Data & downloads

Scroll horizontally to see all columns.

Every municipal record, with zero and missing separate
Reported direction categoryMunicipalities
Two-way, more sent79
Two-way, more received57
Two-way, equal positive counts10
Sent only36
Received only5
Zero in both directions63
Both values missing7

257 municipal services; the first five rows cover 187 positive-flow records. Categories are mutually exclusive.

Scroll horizontally to see all columns.

Weighting and sensitivity among positive-flow municipalities
MeasureAll positive-flow recordsWithout largest gross-flow record
Municipalities187186
Gross directional volume22,23719,530
Volume removed by netting10,67410,674
Gross-weighted cancellation share (%)48.0054.65
Median municipal cancellation share (%)50.0050.00
Absolute normalized balance ≥ 0.509897
Absolute normalized balance ≥ 0.805756

Thresholds are analyst choices. Largest-flow omission is diagnostic; main results retain the record. Cancellation is arithmetic, not reciprocity.

Scroll horizontally to see all columns.

Positive-flow records underlying the scatter plot
Municipal codeSentReceivedGrossNormalized balance
304019215881680−0.890476
20101129411700.517647
403015763120−0.050000
20102242268510−0.050980
50101187250.440000
40101202220.818182
3040264100.200000
30101129471760.465909
3010295531480.283784
10801145−0.600000
40401256691−0.450549
20801151833−0.090909
30501128131410.815603
404024539840.071429
305024566111−0.189189
40403355085−0.176471
105011011.000000
10201055−1.000000
201032091213300.266667
204021011.000000
101012416400.200000
2040397491460.328767
106013140.500000
4030214135149−0.812081
30103207397604−0.314570
20104181582390.514644
40203132205337−0.216617
305036617830.590361
102022816440.272727
3030351015−0.333333
40405198324522−0.241379
30504213152−0.192308
403031851853700.000000
108022626520.000000
102032707027071.000000
1080313001301.000000
3010575261010.485149
30106128187315−0.187302
20602124160.500000
30505123547−0.489362
30202372390.897436
20404318390.589744
20901044−1.000000
10504145−0.600000
501046061.000000
4040674401140.298246
208023938770.012987
30506404989−0.101124
105065051.000000
20202930931.000000
30403222240.833333
30507365288−0.181818
210024083123−0.349593
402065051.000000
404072893121−0.537190
40207110111.000000
304044041.000000
104011011.000000
20105173812540.362205
40304308380.578947
10301450214710.910828
50105254065−0.230769
10507103130.538462
40408213354−0.222222
3040519194213−0.821596
2020421012−0.666667
20405133160.625000
20701286340.647059
2100372681400.028571
204064150.600000
305081510250.200000
207023113440.409091
4010215731600.962500
208046211730.698630
10302175220.545455
20106246410656−0.250000
50106274443180.723270
5010773116189−0.227513
203024435651008−0.121032
501082421604020.203980
30204196250.520000
30107420554750.768421
2040736003601.000000
105082240.000000
205013140.500000
301083360.000000
104033031.000000
20604160161.000000
20805151934−0.117647
10404130131.000000
2010765110.090909
2080692332−0.437500
10103244102540.921260
40409167291960.704082
204082240.000000
108066061.000000
108082130.333333
303045051.000000
10104347−0.142857
108092240.000000
40306255681−0.382716
202063140.500000
303052021.000000
2060591625−0.280000
404105818760.526316
3010928265293−0.808874
2060641822−0.636364
30110157085−0.647059
401035813−0.230769
30205454490.836735
50110424486−0.023256
106034646920.000000
201092532044570.107221
30411268−0.500000
40411202545−0.111111
20110372467839−0.113230
303064041.000000
20410202220.818182
1051202020−1.000000
10105300301.000000
20411180181.000000
10513212230.826087
2060873100.400000
204122021.000000
30111305282−0.268293
208075233850.223529
2030488160.000000
1010691221−0.142857
10107650651.000000
40214122140.714286
40308598670.761194
50111102535−0.428571
104051510250.200000
20305112130.692308
10406100101.000000
403096814−0.142857
403104012520.538462
580084040800.000000
30509122381600.525000
30510203392420.677686
5800905050−1.000000
106024260.333333
3051113321350.970370
401048081.000000
1030412481320.878788
501125124750.360000
1060424502451.000000
10712120121.000000
2061001010−1.000000
580113030600.000000
304074610−0.200000
2070367601270.055118
30307150151.000000
107132312350.314286
4041295561510.258278
205044812−0.333333
303083031.000000
30309102120.666667
201118891179−0.016760
50113106160.250000
30112132671990.326633
403111011.000000
50114142135−0.200000
30408142160.750000
20612270271.000000
304097071.000000
301133311454760.390756
20809119200.100000
20112121461670.449102
10108122739−0.384615
104073031.000000
4031221719−0.789474
4031342135177−0.525424
1010982100.600000
40413220221.000000
103052130.333333
1040864100.200000
107173031.000000
302071011.000000
10110107170.176471
10306235−0.200000
1040921517−0.764706
10719134−0.500000
50116125769−0.652174
102064260.333333
1081351217−0.411765
206142915440.318182

All 187 positive-flow records, identified by source municipal code. Counts are reported directions, not unique network transfers. Ratios rounded only for display.

Evidence & method

How we calculated it

  1. Pair DGLAB RNBP 2024 fields 9.1 and 9.2 across all 257 municipal records; retain zero and keep blanks missing. All 514 selected CSV/XLSX cells agree and marginal totals reconcile.
  2. For positive gross T=S+R, calculate normalized balance B=(S−R)/T and netting cancellation share 2 × min(S,R)/T. Report municipality median and volume-weighted share separately.
  3. Predefined sensitivity uses absolute normalized balance cutoffs 0.50 and 0.80, top 1/5/10 gross-volume shares, and one leave-largest-gross-out calculation. All records remain in main results.
  4. Source inspection preceded the fixed paired-analysis protocol; no externally preregistered or first-ever-method claim. Independent computation reproduced the core raw-CSV results.

What the data cannot tell us

  • One administrative reporting frame and year; no all-Portugal generalization, temporal persistence, causal effect, efficiency or service-quality claim.
  • No document, request or counterparty identifiers. Gross volume can count a transfer at both endpoints; cancellation is not bilateral reciprocity or matched individual documents.
  • Seven paired blanks are unknown, not inactivity. Reported zeros do not prove no service offer; the frame need not form a closed national network.
  • Thresholds are analyst choices, not standards. Large records influence volume-weighted statistics; the influence check is not a justification for removing them.

Original sources

  1. DGLAB RNBP 2024 open dataset

    DGLAB RNBP 2024 open dataset: Official municipal reporting scope, update 14 April 2026 and CC BY 4.0 licence.

    Accessed October 30, 2026

  2. DGLAB RNBP 2024 official CSV

    DGLAB RNBP 2024 official CSV: Fields 9.1 and 9.2; 257 municipal rows. Source blank cells remain missing.

    Accessed October 30, 2026

  3. DGLAB RNBP 2024 statistical report and questionnaire

    DGLAB RNBP 2024 statistical report and questionnaire: Preserved official PDF: municipal scope and response limits; interlibrary totals on printed page 19; questionnaire section 9 on page 85. Fresh endpoint fetch later returned 502; analysis uses the preserved source artifact.

    Accessed October 30, 2026

  4. Creative Commons Attribution 4.0

    Creative Commons Attribution 4.0: Source dataset licence; attribution and modification notice, without endorsement.

    Accessed October 30, 2026

Cite & reuse this study

Alisson Bento. Net lenders can also be borrowers: Portugal’s interlibrary flows. Secondary analysis of DGLAB, Estatística RNBP 2024, municipal interlibrary fields 9.1 and 9.2. Sources accessed October 30, 2026. https://alissonbento.com/research/portugal-interlibrary-netting-and-two-way-flows-2024/

Source data: DGLAB, Estatística RNBP 2024, updated 14 April 2026, CC BY 4.0, https://creativecommons.org/licenses/by/4.0/. Field selection, calculations, labels and original figures by Alisson Bento. No DGLAB endorsement or rights to its logos and report design implied.

Corrections or editorial questions

Pass it on.

Net lenders can also be borrowers: Portugal’s interlibrary flows | Alisson Bentoalissonbento.com

WhatsAppFacebookXTelegramRedditThreadsEmailSMS