Transfer of glass fragments when bottles and drinking glasses are broken

Article Outline

• Abstract
• 1. Introduction
• 2. Materials and Methods
• 3. Results
• 4. Discussion
• Acknowledgements
• Reference
• Copyright

Abstract 

Experiments have been carried out to determine if and how many glass fragments are transferred onto upper garments following breakage of bottles and drinking glasses. In all instances glass was transferred. The numbers of transferred fragments after a bottle is broken ranges from three to twenty five. The numbers of fragments transferred following the breakage of a drinking glass ranges from three to approximately one hundred and twenty. On average three times the amount of glass is transferred following breakage of a drinking glass as compared to breakage of a bottle.

Keywords: Forensic science, Bottle glass breakage, Drinking glass breakage, Glass transfer

Back to Article Outline

1. Introduction 

Bottles and drinking glasses may be broken to create a weapon in the time leading up to an assault. Such glass items may also be broken over the head or into the face of the victim during the course of an assault. There is an expectation that glass will transfer to the perpetrator (and victim) as this happens. One way, therefore, in which a forensic scientist may be asked to assist in the investigation of such an assault is by searching clothing for the presence of glass fragments which could have originated from the broken item. For the results to be of value to the investigator and/or the courts they need to be interpreted within the particular casework context.

An evaluation of the significance of the findings can be carried out using the Bayesian approach to interpretation. This approach to interpretation, the concept of likelihood ratio evaluation, and the limitations of other approaches have been reviewed [1].

Using a Bayesian approach the findings in any case are interpreted in the light of two propositions; for example by considering the “prosecution” and “defence” alternatives.

To evaluate the significance of any glass fragments present on a suspect's clothing within the context of a “glassing” assault the prosecution and defence alternatives may be:

The findings are evaluated by calculating the likelihood ratio given these alternatives.

To assess the likelihood ratio two of the questions which need to be considered are transfer and persistence probabilities, that is:

A number of papers investigating the breakage of different window glass types and glass transfer onto clothing following such breakage has been comprehensively reviewed [1]. My own and colleagues experience of investigating “glassing” assaults indicates that it is possible for glass fragments to be transferred to clothing items when a bottle or a drinking glass is broken. There is no published data, however, on the extent of transfer of such glass fragments onto clothing following the breakage of bottles or drinking glasses. Prior to these experiments interpretation of any findings in which a bottle or drinking glass had been broken had to be inferred from the results of the window glass breakage/ transfer data.

The experiments described were carried out to verify that glass fragments are transferred to clothing when a bottle or drinking glass is broken, and to record the quantities of fragments transferred. An initial set of experiments was carried out to investigate this followed by further experiments to consolidate the data. The data from these experiments is presented in this paper.

The experiments were devised to minimise the second factor in the numerator of the likelihood ratio; that is the loss of glass fragments after transfer. In the experiments all items were taken off and packaged within half an hour of the glass item being broken. Practically, in an actual investigation, the clothing would be removed at time intervals significantly longer that this.

Back to Article Outline

2. Materials and Methods 

A bottle or drinking glass was broken against a brick wall by members of staff wearing different upper garments. The majority of glass items were broken at head height; three were broken at chest height. The garments worn were removed in an area remote from the area of breakage within thirty minutes of the glass item being broken and packaged for later examination. Fleece jackets with long sleeves were worn during the initial studies; tee-shirts and shirts, garments more usually encountered in this type of casework, were worn for the later experiments. Glass fragments were recovered from each item and counted.

All the bottles and glasses broken were empty. The items broken were: six 500ml brown beer bottles, twelve colourless 330ml bottles, six straight sided pint glasses and twelve straight sided half pint glasses (three of which were toughened).

The upper garments worn were either newly purchased or blanked to ensure no glass was present. (The items were blanked by shaking the item as described below and checking to ensure no glass was present in the debris). The pocket on the shirt is a patch pocket with no flaps. The garments were worn over a new disposable scene suit. Hand and face protection was used. A total of nineteen different people, male and female, broke the items.

The glass item was hand held and hit, in all but three instances at head height against an indoor brick wall, until it broke. The remains of the drinking glass or bottle were disposed of. Each participant walked outside the building for a short time before the garment was removed and securely packaged in new packaging. The time between breakage and packaging was no longer than thirty minutes.

Each of the packaged garments was searched in a clean laboratory area. For each garment the work bench was cleaned and covered with new search paper. The outer packaging was brushed down; the debris collected and examined using a stereo zoom microscope with a magnifying capability of up to 50 times. If no glass fragments were found the packaging was considered to be blank. The packaging was opened, the garment removed and then shaken over clean paper. This debris was collected together with the inner bag debris to form the “surface debris”. The debris from inside the breast pocket of the shirt was brushed and collected separately to form the “pocket debris”. The surface and pocket debris was examined using a stereo zoom microscope and the numbers of glass fragments counted. Any fragment numbers of fifty or over are approximate.

Back to Article Outline

3. Results 

The results of these experiments show that in all instances glass fragments were transferred to the upper clothing of the breaker. The numbers of transferred fragments after a bottle is broken ranges from three to twenty five. The data from each of these experiments is presented in Table 1. The numbers of fragments transferred following the breakage of a drinking glass ranges from three to approximately one hundred and twenty. The data from each of these experiments is presented in Table 2.

Table 1. Glass transferred following bottle breakage.

* Bottles broken at chest height

Cells with no entry: No pocket was present in the garment.

Table 2. Glass transferred following drinking glass breakage.

Cells with no entry: No pocket was present in the garment.

Some of the bottles required several blows to break them. The majority of the drinking glasses appeared to “crumple” as they were struck against the brickwork, smashing readily compared to the bottles.

Back to Article Outline

4. Discussion 

The results show that in all instances glass fragments were transferred to the upper garment of the breaker. Hence in a glass transfer investigation in which a drinking glass or bottle is broken in the manor described, it is reasonable to assume that glass fragments are likely to be transferred to the breaker's upper clothing.

The average numbers of fragments transferred following drinking glass breakage is just over three times the numbers of fragments transferred following bottle glass breakage. As these items are of comparable size and were broken in a similar way then it is likely that, on average, a drinking glass produces more glass fragments which then transfer after it shatters as compared to a bottle.

The results show that following the 330ml bottle breakage more glass fragments on average are transferred to the shirt that the tee shirt. This could be because the shirt has long sleeves and hence a greater surface area or that the sleeves of the shirt are closer to the breaking item.

The numbers of glass fragments transferred following breakage of the drinking glasses appears to be much more variable than following bottle breakage. As such detailed analysis of the transfer of fragments from drinking glasses to different garments has not been carried out.

The items were broken against a brick wall. This represents the proximity often encountered in such assaults when the glass item is broken into the victim's face or on the victim's head. The results can therefore be used to assist in advising, strategy setting and interpreting such cases. In particular the data is of use pre-submission in advising the investigator of the potential outcomes of any glass transfer investigations. In a similar way the data impacts on the strategy set once a submission has been made to the laboratory and follows through to assisting in the interpretation of the findings by providing a data set to use in a likelihood ratio calculation.

Breaking an item against a brick wall does not mimic the clothing, skin and hair of the victim. This could potentially affect any glass transfer to the offender by retaining some of the broken glass. Investigation of breakage and glass transfer against a clothed dummy could therefore form the basis of further experiments.

Back to Article Outline

Acknowledgements 

Thanks to staff in the “Violent and Homicide” and “Volume” teams at FSS Chorley for their assistance in the experiments and in reading draft versions of this article.

Back to Article Outline

Reference 

1. Curran JM, Hicks TN, Buckleton JS. Forensic Interpretation of Glass Evidence. Boca Raton: CRC Press LLC; 2000;
   • View In Article