Dial 911… I mean 919! Area code system leads to emergency service misdials

May 21, 2012

Here’s an interesting case study about the large consequences of a seemingly small historical oversight in the US system for assigning telephone area codes.

In Raleigh, North Carolina, the area code is 919, which is similar to the emergency services telephone number of 911.  Until recently within Raleigh, dialling the area code was optional, alleviating the risk of misdialling the emergency services.  However, Raleigh is now large enough that dialling the area code is mandatory which has caused an influx of misdialled calls into the emergency services.

Misdials can be verified in a number of ways – such as at the time of the call, through the operator calling back if the caller hung-up, or through sending out police officers to investigate a hang-up.  The problem has gotten so bad that officers are being sent out to investigate hang-ups every 7.5 minutes on average.

The majority of misdials are caused by the elderly, who are less used to having to dial the area code, and businesses, who often need to dial ‘9’ to get an outside line.  Changing the area code is not really considered an option, as it is felt it will be too complicated.  The Director of Emergency Communications has implored citizens to ‘dial carefully’ – though not likely to be an effective solution.

It’s a fascinating problem that highlights how a lack of human factors foresight can go on to cause major operational issues.

Is it worth remembering this case study within industry, particularly when designing communications systems, controls, procedures, etc., so as to future-proof them?


UK petroleum tanker operations trade dispute: an opportunity to remind ourselves of the human factors aspects of driving operations

May 15, 2012

In many industries and many parts of the world, driving is one of the largest hazards, either because of the nature of the industry, the amount of driving undertaken, the overall standard or culture of driving in that country, or because for the majority of workers driving to and from work is the biggest hazard faced.

Since March 2012 the Unite Union, which represents 90% of fuel tanker drivers in the UK, has been in negotiations with haulage companies over ‘fragmented working practices’ which it claims are affecting not only working conditions but also health and safety.  Details of negotiations and said working conditions are not public knowledge and mainly of concern to those involved in the dispute – but whilst this issue is a ‘hot topic’ it would be a wasted opportunity not to remind ourselves of some of the human factors issues involved in driving operations (and work-related driving) more generally.

Road incidents

It is thought that human failure is a factor in 95% of road incidents.  However a higher percentage of resources go on improving vehicle roadworthiness and construction. Whilst improvements to vehicles can help reduce risk – often by helping the driver ‘control the controllable’ (e.g. utilising ABS braking systems) managing the human and organisational factors affecting driving operations (often out of the direct control of the driver) can be a crucial and effective means to prevent incidents.

For example, “[in the UK] driver sleepiness is estimated to account for around one fifth of accidents on major roads, and is responsible for around 300 deaths per year” (EI Human factors briefing note no. 5).  Fatigue is not something directly controllable; the various factors contributing to fatigue need to be managed instead.  For instance, recognising the added risks presented by long commutes, especially when preceding or succeeding a long shift, in some industries and countries there is a growing trend towards accounting for commuting time when setting shift patterns (for example see Fatigue management for the Western Australian mining industry).

Fatigue is only one potential human factor in road accidents.

A recent study found that a driver’s ability to focus on the driving environment varies depending on the ‘cognitive demand’ placed by non-driving activities. The deeper the level of thought in a driver’s mind, the less he/she focuses on his/her surroundings, and the more likely he/she is to be in an incident.

“Good drivers routinely scan the road ahead and around them, looking for potential hazards that they might need to react to. When drivers face even light levels of cognitive demand, they scan the road less”.  This is why using mobile phones and other devices whilst driving can be detrimental to human performance and illegal in many countries.

A question to pose may be what constitutes ‘cognitive demand’?  Is it just physically distracting activities, like using a mobile phone, or also mental distractions, such as daydreaming, worrying about personal issues, etc.?

Managing the risk of driver distraction comes down to influencing driver behaviour and attitudes.  As ever with occupational human factors, efforts to do so focus on the individual, the job/task and the organisation, through applying consistent rules and standards of working – such as appropriate disciplinary actions for unsafe driving behaviours  (see OGP’s publication on the use of life saving rules) – providing training and procedures, managing the workload effectively, and developing the right organisational culture through sending the right messages to the drivers about what is important.

Hearts and Minds ‘Driving for Excellence’ is a publically available training tool published by the Energy Institute, designed to improve both driver and supervisor hazard awareness and journey planning skills. Containing ready-to-use short training exercises it provides an effective and simple-to-use means of changing driver and supervisor behaviour, improving their appreciation of hazards, and improving the culture around driving operations.

Loading and unloading

Road tanker drivers aren’t just expected to drive but also to load and unload product, whether it be petroleum product or chemicals.  Simple human failures here have the potential to result in complicated (and financially expensive) consequences.  For example:

“A road tanker driver was making a delivery to a customer. One compartment of his tanker was connected to the customer’s diesel tank and was already discharging. The driver then made the connection between the customer’s unleaded tank (tank no. 4) and compartment 4 on his vehicle. He opened the valve and a few moments later realised that he should have connected tank 4 to compartment 3 on the tanker. By then, approximately 600 litres of diesel had been delivered. Almost 3 000 litres of unleaded petrol was contaminated by diesel.” (EI Human factors briefing note no. 15).

Two main factors contributed to this incident: firstly it was found that in this company drivers were paid bonuses based on deliveries made and distance travelled, and so were highly incentivised to meet delivery schedules; secondly, forecourts had to close whilst unloading was taking place, and so forecourt staff often put pressure on drivers to complete unloading quickly.  These factors caused the driver to rush the delivery and not take the time to double-check the connections were correct.

With the intent of understanding incident causation and making improvements, the EI commissioned research and developed Guidance on reducing human failure in petroleum product distribution loading and unloading operations: this explores the human failures behind 22 loading and unloading incidents within three companies, both the immediate causes and root causes.

It found that 1 incident was due to an excessive workload, 2 were due to inconsistency in operations, 5 due to the design of the tasks themselves, 7 due to the design of equipment, and 16, by far the majority, were due to there being a ‘performance culture’ – i.e. getting the job done quickly was considered more important than getting the job done correctly/safely.

This guide provides a series of checklists of contributors to human failure in loading and unloading operations, focusing on:

  • the workplace: for example, are loading areas easy to manoeuvre into and out of? Once parked is there enough space to move around the vehicle? At the delivery point is labelling clear?
  • the tasks: are tasks so routine drivers do them on ‘autopilot’? Are there incentive schemes to make drivers rush jobs? Do drivers need to take shortcuts in the tasks due to time pressure?
  • the selection and attitudes of personnel: is driver training of a high standard? Is driver morale high?
  • the organisation: is there inconsistency in the industry regarding equipment and facilities? Is equipment well-maintained? Is the system to report incidents effective?

Where problems are present, guidance is also provided to solve these.

Managing risk

Driving operations pose a hazard in many industries and there are a number of tools and guidance available to help manage that risk, much of which is widely applicable with a little read-across.

But perhaps it’s worth asking whether managing driving risk is comparable to managing risk in other areas of operation, bearing in mind that, unlike onsite operations, there are a number of variables outside the control of haulier companies – such as the behaviour of other road users, the layout of forecourts, etc.?


New human factors case studies published

April 2, 2012

Some of our readers may recall that Step Change in Safety published Human factors: How to take the first steps…, a practical tool for engaging front line workers in human factors issues – reported on HOF Blog in July 2010.

Step Change in Safety have recently published three more additional case studies for use with the How to take the first steps… guide, freely downloadable from the Step Change website.

The case studies focus on a collision between a vessel and an installation, failure of on-deck winch mountings, and a near miss between a Helideck Assistant and the tail rotor of a helicopter.


OGP publish report on the use of ‘life-saving rules’

March 29, 2012

The International Association of Oil and Gas Producers (OGP) has published a report advocating the use of ‘life-saving rules’ in its member companies.

18 life-saving rules are defined, consisting of 8 ‘core’ rules, and 10 ‘supplemental’ rules.  Some of the rules include:

  • obtain authorisation before entering a confined space;
  • protect yourself against a fall when working from heights;
  • do not walk under a suspended load;
  • wear your seat belt;
  • do not use a phone whilst driving and do not exceed speed limits;
  • work with a valid work permit when required, and
  • verify electrical isolation before beginning work.

The core rules should be applicable to most high hazard industries, with the supplemental rules being suitable for selected industries.

OGP recommends that the life-saving rules be implemented and supported by appropriate disciplinary action for non-compliance.  The report contains a case study on the use of life-saving rules in Shell, where individuals can face disciplinary action up to and including dismissal for non-compliance.  If a rule is broken (even if it didn’t result in incident) Shell first undertakes an investigation to determine why the rule was broken, and then takes appropriate disciplinary action – this may also include investigating supervisors and managers if they were complicit in the rule-break.  Shell is of the view that it is better to take disciplinary action than to risk lives.

The OGP life-saving rules were developed by examining OGP incident data between 1991 and 2010, with the aim that the core rules should cover the direct causes of at least 40% of fatal incidents, and be in line with life-saving rules already developed by OGP member companies.  Together, the core and supplemental rules account for 70% of fatal incidents and 48% of high potential events.

Organisations won’t be expected to implement all 18 rules.  To select the appropriate rules to use, OGP recommends that each organisation undertakes a risk assessment to match the rules up with their own incident data and the type of work they undertake.

Each rule is represented by a symbol based on the ANSI and EU safety icon format. Effort was made to make these as internationally recognisable as possible.

Firm but fair?

Is the use of life-saving rules contentious?  In some cases a single discretion can lead to dismissal, and so opponents have argued that they are too harsh and will lead to a culture of suspicion and blame in the workplace.  Some may also see their use as an alternative to developing safety culture and personal responsibility.

These misgivings may not be justified.

At 2011’s Tripod User Day (September 2011, Amsterdam) Prof. Jop Groeneweg of the University of Leiden gave a talk on the use of life-saving rules in Shell and stressed that they are only worth implementing if they are proven to save lives, otherwise the workforce will not accept them.

The successful use of the life-saving rules relies on non-compliances being reported, which means the rules need to be fully supported and believed in by the workforce.  Consistent enforcement of the life-saving rules may also send the message that the organisation applies consistent standards for safety, and provides the potential to empower front-line workers when taking decisions to stop work on safety grounds.

The OGP life-saving rules themselves focus on preventing only a selection of specific behaviours that are proven to be immediate causes to the majority of fatalities, and should therefore help to standardise the use of life-saving rules across industry.

Implementation

The real issues around the use of life-saving rules are likely to lie in their (mis)application.

In organisations where non-compliances and unsafe working practices are commonplace, or where there is little trust between workers and management, it is questionable whether life-saving rules will be enforceable.  Do organisations therefore require a fairly high safety culture to be already in place before life-saving rules can be implemented?  Implementation of the rules may also depend on the rules themselves being workable – i.e. in there being systems in place to ensure the rules can physically be complied with.

It is important to remember that the life-saving rules mainly relate to fatalities through personal injury (rather than process safety incidents), focus only on strengthening the last line of defence (the worker), and only on a selection of unsafe behaviours.  They are just one part of a larger safety management system, to prevent the failure of a key barrier in what may be a chain of events leading up to an unsafe act, and one means of communicating with the workforce a commitment to safe standards of working.


Changing regulatory oversight of safety management in the US offshore sector

March 22, 2012

The Transportation Research Board (TRB) has released the interim report of the Committee on the Effectiveness of Safety and Environmental Management Systems for Outer Continental Shelf Oil and Gas Operations, which is freely downloadable.

The report explores methods for the US regulators to investigate/audit the implementation and effectiveness of safety and environmental management systems (SEMS) among offshore operators, who should be tasked with investigating/auditing, and the competencies investigators may be expected to have.

The final report is expected to be released after the release of the National Academy of Engineering (NAE) and National Research Council (NRC) report into the likely causes of the Deepwater Horizon incident of 2010.

SEMS

Stemming from the 2010 mandating of API RP 75 by the US Bureau of Ocean Energy Management, Regulation, and Enforcement (BOEMRE), companies operating on the US outer continental shelf are required to have SEMS in place by the end of 2012.  SEMS should work to mitigate human factors root causes in personal safety, major accident hazards (MAHs) and environmental damage.

SEMS will be required to be audited by independent third parties (I3Ps).

In context

The report states that SEMS should operate within four wider organisational processes for managing HSE:

  1. a mechanism for managing HSE (i.e. the SEMS plan);
  2. individual competency;
  3. a strong safety culture, and
  4. personal motivation to behave in the correct way (rewards and consequences).

The implication is that the successful implementation of SEMS cannot simply focus on the SEMS plan itself, but on its use and effectiveness within these wider processes.   This should also be the focus of the auditing and investigation of SEMS.

Investigation methods

The report explores nine possible investigation methods and their advantages and disadvantages, including detailed audits, compliance inspection, peer review-assists, whistle-blower programmes, the use of key performance indicators (KPIs) and mechanical metrics (e.g. from sensors), etc.

Some methods, such as the use of mechanical sensors or KPIs, may be difficult in practice as it is currently unclear which metrics can be used as indicators of an effective SEMS.  Others, such as compliance inspection, may fail to encourage operators to take full ownership and interest in the implementation of SEMS, especially as regulation moves away from prescription-based to risk management-based.

The concern over ownership may also influence who is best to carry out SEMS inspections.  The report states the case for self-inspection, I3P-inspection, regulator inspection, and joint operator/regulator inspection, noting that independent inspections may discourage operators to take full ownership of SEMS, instead delegating the responsibility of performance monitoring on I3Ps.

Competence

The competencies required of SEMS investigators are likely to be broad.  SEMS ruling includes both prescriptive regulations and performance standards, and so SEMS investigation will not only require the reporting of violations of regulations but also identifying weaknesses in the system and opportunities for improvement.

Investigators will also need to be able to focus on both personal safety and process safety, and it is likely that a team of investigators with a number of competencies will be required to complete an audit/investigation.

Lastly investigators will need training and certification.

Commentary

This report represents the initial findings of the committee, but gives an indication of how the US offshore sector may be regulated in the next few years.

Whilst the committee has reviewed various inspection methods in light of how well they encourage operators to take ownership of the process, it is perhaps worth noting that not all inspection methods explored in the report will be suitable for all organisations, as the effectiveness of methods may depend on the safety culture of the organisation.

The 2010 OGP publication A guide to selecting appropriate tools to improve HSE culture provides some guidance on the types of HSE tools that are appropriate for different cultural levels.  It suggests that some methods may not be practical within more reactive cultures, whilst other methods may not allow organisations with more advanced cultures to take full ownership of the investigation process.  For example, peer assists are only really appropriate for ‘proactive’ and ‘generative’ cultures (the two highest cultures on the safety culture ladder).

If the culture of the US offshore sector is to adapt from a prescriptive regulatory framework to a broader approach to risk management using a variety of methods, do operators therefore need to consider their own safety cultures when selecting appropriate investigation methods?  Are a variety of methodologies needed?  Could there be a danger if the methodologies advocated are too narrowly defined?


Natural disasters make human errors more likely

February 23, 2012

Researchers have found that humans are more prone to making mistakes in the aftermath of natural disasters.

Whilst previous research suggests this happens after man-made disasters – for example, leading to an increased number of fatal car accidents – no research had before looked at natural disasters until the 2010 earthquake that hit Christchurch, New Zealand, presented researchers with an unexpected opportunity.

Participants who became anxious after the disaster displayed increased reaction times than normal, but made a larger number of mistakes.  Those who became depressed had slower reaction times than normal.

“These findings also suggest that police, emergency responders, and others working in the aftermath of the disaster may also experience cognitive disruption, which can interfere with their ability to perform rescue-related tasks.”

Although based on a single study, these findings may be relevant to those working in the energy and allied process industries, particularly when planning emergency responses, however further research may be needed.


Human factors – the biggest challenge in offshore oil and gas safety?

February 21, 2012

A survey of offshore safety professionals, conducted in advance of the 2nd annual Offshore Safety Summit on 19-21 March 2012 in Aberdeen, has reported that 48.6% of respondents see human factors as the biggest challenge to the offshore sector at the moment.

This is compared to 10.8% who see problems with technology and equipment as the key issue, and a similar amount who believe having the right HSE processes in place and reacting to new legislation to be the biggest challenges.

Perhaps not surprisingly, nearly 75% of respondents believe that since the 2010 Gulf of Mexico incident there is more pressure being placed on industry by the regulators to improve health and safety.

This raises several questions.

There is a wealth of information available on human factors, including the EI website, HSE website, OGP and Step Change in Safety websites, and more on the way.  Does the challenge of human factors arise from there not being enough information available? (If so, what more is needed?)  Does it stem from a lack of understanding of human factors?  (What can be done?)  Is it because human factors is such a departure from the hard science and technology subjects that engineers and operators usually know so well?  Or is human factors simply a difficult subject to integrate into an organisation’s safety management system?


Human factors developments for 2012

February 8, 2012

In the last HOF Blog post, you learnt about the HOFCOM, the EI’s Human and Organisational Factors Committee, who oversee development of our human factors resources – most of which are available freely online from the EI website – www.energyinst.org/humanfactors.

The HOFCOM undertakes a number of human factors projects every year, producing new resources like guidance for industry and practical tools, and it should be no surprise that 2012 will be no exception!

Training materials to bridge the gap

Work is underway to develop training materials for a 4-day ‘human factors foundation course’.

Currently there is a gap in knowledge between those who have an awareness of human factors issues and those who are able to fully run and manage human factors initiatives, making it difficult for companies to train people up to the required standard to oversee large projects and to manage human factors in their own work.  One reason for this is that human factors is not always adequately addressed within degree level engineering courses.

This project aims to develop a complete set of training materials for a 4-day course, including a slide pack, teaching notes, workshop exercises and assessment questions.  Negating the need to produce bespoke materials for every new human factors training course, these materials will be made available to companies/institutions, either to run a stand-alone course, or to incorporate into existing courses, including at undergraduate and masters level.

The materials are being developed by Human Reliability Associates, who run the human factors module within the University of Sheffield’s ‘Process safety and loss prevention’ MSc(Eng) course, and are being developed in consultation with IMechE, Cogent, HSE and HSL, among other stakeholders.  The goal is for these materials to set the standard for human factors training.

Web-based course

In support of this, the EI is soon to launch a free web-based Human factors awareness training course.  This aims to give those working in industry an awareness of the most pertinent human factors issues and some of the methodologies by which they can be managed.

Developed by Greenstreet Berman, the course also assesses participants’ learning and records their score, intending to raise awareness and prepare for further development, such as more advanced training as provided by the EI’s human factors and accident investigation face-to-face training courses and the foundation course materials.

Qualitative human reliability analysis

Directed by the EI’s SILs/LOPA Working Group, we are producing Guidance on quantified human reliability analysis (QHRA).

Integration of human factors into major hazards operations subject to safety cases/reports requires use of robust data in risk assessments; however, there lacks practical information on the use of human error probability (HEP) data in human reliability analysis (HRA). This project aims to develop practical guidance on quantification of human failure in risk assessment for a primary audience who need to be competent ‘intelligent customers’, whether such services are provided in-house or using external resource.

Fatigue management and crew resource management

The HOFCOM have also received funding for two more projects to begin this year.

Firstly to revise the EI fatigue guide: Improving alertness through effective fatigue management.  First published in 2006, this guidance document will be updated with the latest good practice on fatigue management, including on the growing area of fatigue management/recording systems, as well as pertinent new research findings.

Secondly, we aim to produce a guidance document on crew resource management (CRM) and non-technical skills (notechs).  This new guidance document will set out the case for CRM/notechs training, its benefits, its subject matter, and good practice on what CRM/notechs training should include.  The UK HSE believes there is a strong case for the energy sector to adopt CRM principles and this document aims to highlight this.

Hearts and Minds

Last but not least, the EI co-ordinated Hearts and Minds programme has funded two PhDs on ‘learning from incidents’ and ‘safety leadership’ respectively, both of which are nearing completion.

It is anticipated that the findings of this research will drive further development of the Hearts and Minds toolkit (watch this space), including plans to conduct follow-up research and to produce a practical ‘learning from incidents’ toolkit to help sites engage with and apply learning in their local operations.

Phew! If that’s 2012, what might 2013 look like?

What would you like to see produced to help you manage human factors?  Perhaps you identify with some of the gaps raised here?  Or maybe you’d like to share different views on what you might like to see in HOFCOM’s forward work programme, based on what would be helpful to your operations?

To keep up to date with the EI’s human factors work programme, visit the EI human factors website.


Who are the HOFCOM?

February 3, 2012

You’ve heard of the HOF Blog (after all, you’re reading it!) but who are the HOFCOM?

The EI offers a strong human factors technical work programme with a focus on the energy industry, making a large collection of human factors resources (tools, guidance documents, etc.) available freely on its website, www.energyinst.org/humanfactors.  Furthermore, we’ve been adding to that collection by 2 or 3 titles per year, developed training courses and ran conferences.

But have you ever wondered who makes it all happen?

The Human and Organisational Factors Committee (HOFCOM) was formed in 2001.  Comprising representatives from BP, Shell, ConocoPhillips, Magnox Sites, the Health and Safety Executive, and specialist consultants, all giving their time freely, the HOFCOM steers the EI’s human factors work programme.

Meeting 4 times per year, the HOFCOM is responsible for proposing new human factors projects for funding by the EI’s Technical Partners.  Ideas for projects may arise from issues encountered within committee member’s own companies, current ‘hot topics’, regulatory concerns, or in response to incidents.  While committee members draw on their own experiences and those of their respective companies, they leave their ‘corporate hats’ at the door – ensuring proposed projects are for the benefit of all industry rather than just one particular company.

The HOFCOM will define the project and then oversee its development by a subject specialist contractor.  After carefully selecting a contractor to develop the work, the HOFCOM then steers the project from start to finish.  In the process, they will call upon the help of industry and subject matter specialists to peer review work, the goal being to ensure that the finished product meets a high standard and the needs of the energy industry, whether operating companies, regulators or consultancies.

The result?  In just the last few years alone a wealth of new resources has been produced, for example:

So there you have it, now you know who the HOFCOM are and what they do.  The question is: how could you become engaged?  This could be through reviewing work, highlighting key areas for future development, or, if you are an employee of one of the EI’s Technical Partner Companies or its Technical Company Members for the safety theme, you might want to consider becoming a member of the HOFCOM.  Contact Stuart King sking@energyinst.org for more information.


The next rung on the ladder in staffing arrangements – does HSE CRR348/2001 need updating?

January 27, 2012

In 2001, the UK HSE published Contract Research Report (CRR) 348/2001 Assessing the safety of staffing arrangements for process operations in the chemical and allied industries.  This report sets out a 2-part methodology for assessing whether staffing levels and arrangements are adequate to maintain the safety of operations in high hazard industries, using:

  • Physical assessments – which assess the ability of staff to successfully detect, diagnose and recover hazardous scenarios.
  • Ladder assessments – which benchmark organisational factors in relation to industry best practice.  CRR348/2001 provides ladders on eleven staffing issues.

In 2004, the Energy Institute launched a guidance document setting out a best practice approach to the CRR348/2001 methodology that captures learnings from its use, and provides an additional ladder for the use of automated systems.  This guidance document was authored by Dr. Andy Brazier, now an independent consultant.

A recent independent paper by Andy Brazier provides a ten year review of the CRR348/2001 methodology.  It offers some personal observations of common human factors issues still prevalent within many organisations and makes some recommendations.

Of particular interest, the paper also suggests an improvement to the ladder assessments provided in CRR348/2001.

Ladder assessments can be described as a hierarchy of statements and keywords that describe typical practice on how organisations manage (or fail to manage) particular activities, in this case those associated with managing staffing arrangements (including situational awareness, team working, alertness and fatigue, etc.).  These make up the ‘rungs’ on the ladder, with each successive rung representing better practice than the last.

As well as a method to benchmark performance, ladder assessments provide a series of checkpoints for improvement, with the top rung representing ‘best practice’.  Teams must be confident that they fully meet the expectations of one rung before they can meet the next one up.  As such, CRR348/2001 provides an overview of good and best practice (as of 2001) and a route to achieve that.

However, Andy makes the suggestion that the top rungs in these ladder assessments no longer represent best practice – “I believe the concept of High Reliability Organisations, that has developed over the last decade, gives us a good reference point for what may be considered as best practice now” – and proceeds to suggest an additional rung for each ladder detailing practices commonly found in high reliability organisations (HROs).

Does this mean CRR348/2001 is now out of date or no longer valid?

This is unlikely.  In 2009 the EI’s Human and Organisational Factors Committee (HOFCOM) reviewed and reaffirmed the EI guidance document, believing it (and the CRR348/2001 methodology it supplements) to still be valid and useful.

When using ladder assessments, we should remember that companies must fully meet the requirements of the rung below before they can move up, so until they have reached the top rung CRR348/2001 is still useful.  However, for those companies approaching the top rung, or with serious ambitions of becoming a HRO, would adding an additional rung to the ladder be beneficial?


Follow

Get every new post delivered to your Inbox.

Join 54 other followers